Volume 13, No. 2 Spring 2000 • Official Publication of the ...ies evolving from theoretical models...

Volume 13, No. 2 Spring 2000

• Official Publication of the Mid-Western Educational Research Association •

The University of Akron

Information for Contributors to the Mid-Western Educational Researcher

The Mid-Western Educational Researcher accepts research-based manuscripts that would appeal to a wide range of readers. All materials submitted for publication must conform to the language, style, and format of the

Publication Manual of the American Psychological Association, 4th ed., 1994 (available from Order Department, American Psychological Association, P.O. Box 2710, Hyattsville, MD 20784).

Four copies of the manuscript should be submitted typed double-spaced (including quotations and references) on 81/2 x 11 paper. Only words to be italicized should be underlined. Abbreviations and acronyms should be spelled out when first

mentioned. Pages should be numbered consecutively, beginning with the page after the title page. Manuscripts should be less than 20 pages long. An abstract of less than 100 words should accompany the manuscript.

The manuscript will receive blind review from at least two professionals with expertise in the area of the manuscript. The author's name, affiliation, mailing address, telephone number, e-mail address (if available), should appear on the title page only.

Efforts will be made to keep the review process to less than four months. The editors reserve the right to make minor changes in order to produce a concise and clear article.

The authors will be consulted if any major changes are necessary.

Manuscripts should be sent with a cover letter to:

Mary K. Bendixen-Noe, MWER Co-Editor 1179 University Dr., The Ohio State University at Newark, Newark, OH 43055

The Mid-Western Educational Researcher (ISSN 1056-3997) is published quarterly by the MidWestern Educational Research Association through The Ohio State University. The Summer issue serves as the annual meeting program. Non-profit postage paid at Columbus, Ohio, with permission of the College of Education, Daryl Siedentop, Interim Dean.

POSTMASTER: Send address change to Jean W. Pierce, Dept. EPCSE, Northern Illinois University, DeKalb, IL 60115.

On the Cover With an enrollment of more than 23,300, The University of Akron has the third-largest main campus

enrollment of Ohio’s state universities. A major teaching and research institution, it is known worldwide for its work in polymer science and polymer engineering, and in many other areas—including engineering, organizational psychology and educational technology.

Accredited since 1914 by the North Central Association of Colleges and Schools, the University offers more than 200 undergraduate majors and areas of study, leading to associate and bachelor’s degrees. For advanced study, the University provides more than 100 master’s degree programs and options, 17 doctoral degree programs and four law degree programs.

The University of Akron is home to the world’s first College of Polymer Science and Polymer Engi-neering. What began nearly 80 years ago as a few courses in rubber chemistry has evolved into the largest academic program of its kind in the world. The University also is the only university in Ohio to have a basic science graduate program ranked among the top five nationally.

The University of Akron is currently being led by its 15th president, Dr. Luis M. Proenza, who began his tenure on Jan. 1, 1999. Under his leadership, the University is pursuing an ambitious program to simul-taneously expand research and enhance teaching excellence.

A $200 million transformation of the University will begin in a few months, creating a “new landscape for learning.” The campus remake will begin in spring 2000 with simultaneous projects totaling more than $24 million, including construction of a new Arts and Sciences Building and an annex to the Sidney L. Olson Research Center, which houses the department of polymer engineering. Both projects should be completed in 2001.

Volume 13, Number 2 · Spring 2000 Mid-Western Educational Researcher 1

MWER Publication Address

Mary K. Bendixen-Noe The Ohio State University, Newark

1179 University Dr. Newark, OH 43055

Phone: (740) 366-9469 Fax: (740) 366-5047

email: [email protected]

MWER Subscription & MWERA Membership Information

Jean W. Pierce Dept. EPCSE

Northern Illinois University DeKalb, IL 60115

Phone: (815) 753-8470 Fax: (815) 753-9250


Editorial Advisory Board

Thomas Andre Iowa State University

Josue Cruz University of South Florida

Louise Fleming Ashland University

Tom Ganser University of Wisconsin–Whitewater

Kenneth A. Kiewra University of Nebraska, Lincoln

Isadore Newman University of Akron

A. William Place University of Dayton

Rose Mary Scott University of Wisconsin at Parkside

Co-Editors

Mary K. Bendixen-Noe The Ohio State University,

Newark email: [email protected]

Kim Metcalf Indiana University


Layout and design

Judy E. Teska JET Desktop Solutions

Volume 13, Number 2 Spring 2000 ISSN 1056-3997

Consulting Editors

Deborah L. Bainer The Ohio State University,

Mansfield

Gene A. Kramer American Dental Association

Richard M. Smith

Development of Attitude Toward Teaching Career Fred L. Pigge and Ronald N. Marso, Bowling Green State University

2

Using Action Research to Facilitate School Improvement Susan N. Kushner and Sharon D. Kruse, University of Akron

10

Annual Meeting Making a Difference in the 21st Century

15

Reading Ability as a Predictor of Student Success in Business School

Robert E. Pritchard and George C. Romeo, Rowan University 23

Culture and Strategy in Business Schools: Links to Organizational Effectiveness Christopher Clott, Elmhurst College

Nancy Fjortoft, Midwestern University 27

Traditional vs. “Resampling” Approaches to Statistical Inferences Regarding Correlation Coefficients

Thomas R. Knapp, Gerald L. Noblitt, and Sunanta Viragoontavan The Ohio State University

34

Mid-Western Educational Researcher Volume 13, Number 2 · Spring 2000 2

The continued emphasis upon accountability and pupil competence in public education increasingly raises concern about teacher quality and development. This concern has resulted in several recent efforts to consolidate and reassess existing theory and research on teacher development. For example, Firestone and Pennell (1993) reviewed approxi-mately 30 studies investigating relationships between work-ing conditions and the development of commitment to the teaching profession, Brookhart and Freeman (1992) analyzed more than 40 studies of the nature of entering teachers for common research themes and procedures, and Kagan (1992) examined approximately 50 studies of preservice and be-ginning teachers for relationships with various theoretical models of teacher development.

Among the common themes noted in the aforementioned integrative interpretations of the existing teacher develop-ment research literature were an increasing number of stud-ies evolving from theoretical models of teacher development, such as Fuller’s (Fuller and Bown, 1975) model of the de-velopment of teaching concerns and Berliner’s (1988) model of teacher cognitive development; a recognized need for, but relative absence of, longitudinal studies of teacher de-velopment; and the growing acceptance of the presence and understanding of stages in teacher career development. Within these studies; teacher career development was viewed as changes in teachers’ job skills, knowledge, and behav-iors; changes in attitude and outlooks; and changes in job events (Burden, 1982). Evidence from these studies indi-cated that these changes follow a developmental pattern and interact with teachers’ beliefs, prior experiences, and per-sonality (Bendixen-Noe and Redick, 1995; Kagan, 1992).

In the present study, attitude toward teaching as a ca-reer was investigated using a sample of teacher candidates as they progressed through teacher preparation and the early years of their classroom teaching. Research on attitude de-velopment in teacher candidates suggests a pattern of change from early formalized and rigid attitudes toward teaching to a more liberal, democratic, and humanistic attitude about

teaching in mid teacher preparation and returning to the former rigid control of pupils orientation following student teaching and early teaching responsibilities (Callahan, 1980; Hogben and Lawson, 1984; Hoy and Woolfolk, 1990; Lipka and Garlet, 1981). Hoy and Rees (1977) suggest that this regression to an earlier attitudinal position following early teaching experiences may simply result from the teacher candidates’ return to the conformity of the conservative, bureaucratic behavior norms of the public schools. In con-trast to these research findings, some cross-sectional stud-ies of preservice and inservice teachers have revealed stable attitudes, or even positive changes, rather than negative changes in attitude toward teaching during the transition from students to teachers (Marso and Pigge, 1989; Paschal and Treloar, 1979; Sandgren and Schmidt, 1956). The danger inherent in these studies is that many teacher candidates who had experienced an abrupt decline in attitude toward teach-ing may have already left the profession before the mea-surement points and, therefore, do not appear in the later inservice samples of teachers.

The aforementioned discrepancies in research findings may be explained by suggestions that attitude development, like more general teacher development, is influenced by prior personal experiences and individual characteristics of the prospective teachers and by characteristics of the employ-ing school (Tabachnick and Zeichner, 1984; Zeichner, 1980). Some empirical evidence exists which supports this expla-nation. For example, Villeme and Hall (1980) and Pigge and Marso (1987) reported that teacher candidate gender and grade level of instruction are related to their attitude toward teaching. Also, Byler and Byler (1984) revealed relation-ships between extent of student teachers’ prior field experi-ence, student teachers’ own expectations of their experience, and characteristics of cooperating teachers and the attitudes and morale of student teachers. Regarding the extent of field experience factor and its possible impact upon teacher can-didate attitudes, Gibson and Coleman (1997) reported that most schools of education have increased field requirements

Development of Attitude Toward Teaching Career Fred L. Pigge and Ronald N. Marso

Bowling Green State University

Abstract The purpose of this longitudinal study was to investigate whether selected academic ability indices and personal characteristics of 117 teachers were associated with changes in their attitude toward teaching as a career measured upon the commencement of teacher preparation, the end of student teaching, and upon completion of their fifth year of teaching. Mixed model two-way ANOVA procedures revealed that the attitude toward teaching of the neophyte teachers remained stable and positive during teacher prepa-ration but were less positive near the end of their fifth year of teaching; that ACT scores, Rotter’s locus of control, Myers-Briggs Type Indicator preferences, gender, when the decision to teach was made, and the initial degree of assurance about becoming teachers were associated with attitude toward teaching as a career; and statistical interactions were identified, revealing contrasting changes in attitude devel-opment during teacher preparation for candidates with different personal characteristics.


for teacher candidates. This change in itself may account for some of the differences in the results of the studies con-ducted at different times.

Relative to the broader question about the importance of teacher attitudes, relationships have been found between teacher attitudes and success in the transition from students to teachers as well as between attitudes and classroom teach-ing behaviors, Villeme and Hall (1980) reported that candi-dates with more positive attitudes toward teaching and with higher grade point averages were more likely to actually enter the teaching field; Ramsay and Rensley (1986) found that teacher attitudes influenced the achievement of their pupils; Sorenson, Schaefer, and Nyman (1966) reported that teacher attitude was related to teacher persistence in the profession; Bunting (1988) noted a relationship between teacher atti-tudes and choice of instructional procedures; Friesen, Prokop, and Sarros (1988) reported a relationship between negative attitudes (e.g., depersonalization) and teacher burn-out; and Noad (1979) identified a relationship between teacher attitudes and their pupils’ ratings of their teaching performance. Additionally, Wilkinson (1994) states that newly qualified teachers have an “alarmingly high” attrition rate; Valli (1992) indicated that beginning teachers experi-ence more classroom problems and experience them more profoundly than do experienced teachers; and Chester and Beaudin (1996) state that it is “well documented” that teach-ers’ self-efficacy beliefs and attitudes toward teaching change during their first year of teaching.

In summation, the developmental perspective of teacher evolution suggests that attitude toward teaching will change during teacher preparation and the initial years of teaching, and research findings derived from cross-sectional samples indicate that teacher attitudes may be related’ to both teacher attrition and the performance of their pupils. The research literature also suggests that various academic and personal attributes of teacher candidates, as well as the nature of the teaching setting itself, may interact with the development of teacher attitudes The purpose of the present study was to further investigate selected academic ability and personal characteristics of teachers that may be associated with changes in attitude toward teaching for a longitudinal sample of teacher candidates as they progressed through prepara-tion and the initial years of teaching

Method

The sample for this study was comprised of 117 neo-phyte teachers who completed the Attitude Toward Teach-ing as a Career measure at the commencement of teacher preparation, near the end of their student teaching practicum, and again near the end of their fifth year of teaching. This 11-item attitude scale provides a single score responded to on a continuum from strongly disagree ‘1’ to strongly agree ‘6’; thus, a score of 66 represents the maximum positive attitude score from this instrument. This instrument is unique in the measurement field in that it provides discriminant

validity related to career choice, and more recent research (Pigge and Marso, 1992) has indicated that the scale differ-entiates between teacher candidates persisting or not per-sisting through teacher preparation. It is a situation specific scale based upon need theory and the attitude-concept view of attitude structure. Scores from the scale are conceptual-ized as a function of the individual-belief value matrix wherein attitudes evolve from perceptions that the attitude objects block (negative) or facilitate (positive) need satis-faction. From a teacher development perspective, one could hypothesize that teacher candidates should show an increas-ingly positive attitude toward teaching as a career as their knowledge and skills develop during teacher preparation and that this positive attitude should remain stable during the initial years of teaching.

The following data also were gathered from the sub-jects during teacher preparation: American College Test (ACT) and Comprehensive Test of Basic Skills (CTBS) com-posite scores, university and education grade point averages, Rotter’s (1966) externality locus of control scores, univer-sity supervisors’ ratings of their student teaching perfor-mance, and Myers-Briggs Type Indicator personal preference classifications (Myers and McCaulley, 1985). The CTBS is one of the most frequently used K-14 grades standardized achievement batteries with a focus upon assessment of read-ing skills, language acquisition, and mathematical compu-tations and concepts. The ACT was developed by the American College Testing Program, and it has been one of the two major undergraduate college admissions tests used in the country over the past three decades. The Rotter and Myers-Briggs instruments have been heavily used for vari-ous research purposes over a period of many years, and both have been found to be related to various teacher character-istics. Similarly, the gender, grade level of instruction, and time when the decision to teach was made classifications have been found to be related to teacher persistence (Marso and Pigge, 1997) and have been suggested as factors to be considered in recruiting teacher candidates (Hutchinson and Johnson, 1993-94).

The researcher developed university supervisors’ scale for rating student teachers’ performance encompasses as-sessments in six performance categories: presents content effectively; effectively plans, prepares, and organizes instruc-tional activities; maintains a positive learning climate; main-tains appropriate student behavior; displays professional knowledge and behavior; and shows fairness, tact, compas-sion, and good judgment. Each of these performance cat-egories is accompanied by behavior descriptions of the types of student teachers’ performance to be assessed. The six items are responded to on an eight-point scale from the worst ‘0’ to truly exceptional 7, yielding a total score from zero to 42 The scale directs the university supervisor to rate the current student teacher relative to the typical performance of all student teachers supervised over the prior five years.

This sample of teacher candidates was beginning its teacher preparation at a large Midwestern teacher prepara-


tion institution. The candidates were predominately white (98%), female (81%), elementary (57%) and secondary (43%) majors, very certain or almost certain about teaching (88%), and from rural (33%) or suburban (54%) high schools of mod-erate to small size (61% with high school graduating classes of 300 or less). Most of the candidates were employed in schools similar to those from which they had graduated.

Mixed-model two-factor ANOVA procedures with one repeated measures factor were used to analyze the data. The three points of time in teacher development, prior to and following teacher preparation and at the end of the fifth year of teaching, comprised the repeated measures factor inde-pendent variable. The various academic ability indices and personal attribute classifications of the teachers were used as the second independent variable. In all analyses, the atti-tude scores were used as the dependent variable.

The specific row classifications used in the 3x2 and 3x3 ANOVA procedures for the academic ability and personal attribute classifications were approximate high and low halves of the academic ability indices derived from the ACT and CTBS composite scores, the university and education grade point averages, and the student teaching performance ratings; the dichotomous personal classifications of gender, grade level of instruction (elementary and secondary), Myers-Briggs’ personal preference types (extraversion-in-troversion, sensing-intuition, thinking-feeling, and judging- perceptive); and three level classifications from Rotter’s locus of control (approximate top, mid, and low one-thirds of the externality scores) and from when the teacher candi-dates first decided to become teachers (prior to, during, or following the high school years).

Results

The results of the ANOVA procedures indicated that the teachers’ attitude toward teaching as a career changed during this period of teacher development. As shown in Table 1, the overall attitude means for these three points in career development were almost identical from pre- to post-teacher preparation, but by the fifth year after graduation the teach-ers’ attitude toward teaching as a career had become less positive (F = 5.06, p = .0071). The series of ANOVAs also revealed significant row main effect or interaction F ratios for 6 of the 15 subject academic and personal classifica-tions. It can also be observed in Table 1 that the time vari-able was significant each of the six comparisons presented (p’s from .0004 to .0122). Post-hoc pair-wise mean com-parisons via a Scheffe test (α.10) revealed no significant differences between the sophomore and senior means, but both of these sets of means were higher than the means after the fifth year of teaching.

Assurance About Teaching Classification

The assurance classification revealed an overall attitude mean difference among the very certain, certain, and uncer-tain about teaching candidates, F = 5.82, p = .0039 but re-

vealed a nonsignificant assurance χ time interaction as shown in Table 1. The post-hoc pair-wise mean comparisons indi-cated no difference between the certain and the other two groups of candidates, but the uncertain candidates differed from the very certain candidates. Those teachers who were more assured of their decision to become teachers upon com-mencement of preparation reported more positive attitudes toward teaching at all three measurement points as can be seen in Table 2. It can also be seen in Table 2 that the stan-dard deviations for the attitude scores for the assurance, and all other significant classifications, were most diverse at the end of the fifth year of teaching. This increased diversity in attitude toward teaching suggests that the classroom teach-ing experience may have been satisfying the needs of some candidates but likely not for other candidates. It also can be noted that attitude diversity was greater within the groups of teachers who had reported being uncertain (SD = 10.47) and certain about teaching (SD = 9.03) as compared to those who had reported being very certain (SD = 7.94) about teach-ing upon commencement of teacher preparation.

Locus of Control Classification

The locus of control main effect did not reach statisti-cal significance (F = 2.05, p = .1364)11 however, the locus of control χ time in career interaction effect was significant (F = 2.64, p = .0359) as reported in Table 1. The pattern of these means and accompanying effect sizes as shown in Table 2 and Figure 1 indicates that the high externality teachers, those teachers feeling they had little control over their envi-ronment, reported less positive pre- and post-preparation attitudes but somewhat more positive attitudes after teach-ing as compared to the low and mid externality teachers. The low and mid externality candidates reported higher atti-tudes during teacher preparation followed by declines in positive attitude from end of preparation to the fifth year of teaching (ESs = 1.10 and .49, respectively, see Figure 1) while just the opposite occurred for the high externality group (ESs = .27), Typically, effect sizes (ES) of .20 to .50, .50 to .80, and .80 plus are considered small, medium, and large.

As can be seen in Figure 1, the decline in attitude to-ward teaching was particularly sharp for the low externality (ESs = 1.10), those internally oriented teachers who felt they had considerable control over their environment, candidates. This interaction suggests that the high externals found’ class-room teaching to be a more need satisfying experience than what was expected; whereas the low externality candidates in particular found their early teaching experiences to be much less satisfying than anticipated. Perhaps these inter-nally controlled teachers found that they, themselves, could not control the teaching setting as much as they had antici-pated. This finding may be particularly significant as other researchers have reported that internally controlled teach-ers, as might be expected by definition of the internal orien-tation, feel more responsible for the progress of the pupils (Ashton, Webb, and Doda, 1983) and have pupils who achieve higher than pupils of external teachers (Murray and


Tabl

e 1

2x3

and

3x3

ANO

VA F

Val

ues f

or th

e At

titud

e To

war

d Te

achi

ng a

s a C

aree

r Mea

ns a

t Thr

ee T

imes

in T

each

er D

evel

opm

ent a

nd fo

r Sev

en C

lass

ifica

tions

of T

each

ers


Staebler, 1974). Consequently, this interaction may suggest that the transition from students to teachers may be most difficult attitudinally for those teachers initially most posi-tive about becoming teachers and who become most con-cerned about the progress of their pupils.

Academic Ability Classifications

The CTBS, university and education grade point aver-ages, and the student teaching performance row classifica-tions of the teacher candidates did not result in statistically significant main or interaction effects. The ACT classifica-tion, however, revealed a significant main effect mean dif-ference but a nonsignificant interaction effect. The high ACT

candidates reported a more positive attitude toward teach-ing than did the low ACT candidates, F = 4.36, p = .0401 as shown in Table 1. The attitude means at the three career points, as reported in Table 2, indicate that the difference in attitude between the high and low ACT groups was barely evident at the commencement of preparation but was more evident both at the end of teacher preparation and the fifth year of teaching. Both the high and low ACT candidates reported their least positive attitudes at the end of the fifth year of teaching. The low ACT candidates, however, re-ported somewhat less positive attitudes at all three career points and reported the larger decline in attitude over the seven-year period.

Table 2 Means, Standard Deviations, and Ns for Classifications of the Teachers Revealing Differences in Attitude Toward Teaching as a Career at the Three Times in Development


Myers-Briggs Preference Classifications

Three of the four Myers-Briggs classifications resulted in significant main, effect attitude mean’ differences, but none of the time in career χ Myers-Briggs classification interac-tions were significant (see Table 1). The perceptive candi-dates, those preferring a flexible and spontaneous way of life, reported a less positive attitude than the judging candidates who prefer a more planned and orderly way of life, F = 7.86, p = .0061. The extroverted candidates, those preferring the outer world of people and things, reported a more positive attitude toward teaching than the introverted candidates, those preferring the inner world of ideas, F = 6.22, p = .0142. And the sensing candidates, those preferring to work with known facts, reported a more positive attitude about teaching than the intuitive candidates who prefer to work with possibilities and relationships, F = 4.93, p = .0287 as shown in Table 1. As noted earlier, the standard deviations for the attitude scores after the fifth year of teaching were larger than at the two earlier career points.

It can be noted from Table 2 that the majority of these candidates were classified as judging and extroverted rather than perceptive and introverted, but the candidates were rather evenly divided within the sensing-intuitive classifica-tion. The nonsignificant interaction effects identified in Table 1 suggest that none of these preference classifications of the teachers revealed variations in attitude other than the previ-ously reported small general decline from the end of prepa-ration to the end of the fifth year and even though greater diversity of attitudes was reported after teaching as com-pared to the two preservice measurement points.

Summary and Discussion

Differences in attitude toward teaching as a career were identified across the three career development points in this

longitudinal study of 117 teacher candidates. Attitudes to-ward teaching remained stable and positive between the com-mencement of teacher preparation land the end of the student teaching practicum, but attitudes toward teaching became less positive between the completion of the student teaching practicum and the fifth year of teaching. Also considerably more diversity in attitude was noted at the fifth year of teach-ing than at the two earlier career points. The finding of no gain in positiveness of attitude during teacher preparation and the decline in positiveness of attitude from the end of teacher preparation to the fifth year of teaching is contrary to the presumptions underlying the attitude measure and re-lated developmental theory. Attitudinal theory models sug-gest that attitudes toward teaching should become more positive as novice teachers develop their professional knowl-edge and teaching skills and as they find that teaching satis-fies their needs.

Just the ACT classification among the selected academic ability indices was found to be related to the teacher candi-dates’ attitude toward teaching. The teacher candidates with higher ACT scores reported higher levels of attitude toward teaching than did their cohorts with lower ACT scores. The CTBS scores, university and education grade point aver-ages, and student teaching performance ratings classifica-tions were not found to be related to the candidates’ attitude toward teaching. This lent scant support for the question of a relationship between attitude development and academic performance indices of teacher candidates.

Among the personal classifications of the teacher can-didates, the assurance about teaching, locus of control ori-entation, and the Myers-Briggs preference classifications were found to be related to the novice teachers’ attitude to-ward teaching as a career providing somewhat more sup-port for the question of a relationship between attitude development and candidate personal traits. The gender, sec-ondary or elementary school major, time of decision to teach, and presence or absence of teachers in the family classifica-tions were not found to be related to the teaching candi-dates’ attitude toward teaching. Generally, these findings related to candidate characteristics indicated that those more confident about the decision to teach, those feeling they can influence their world (an internally rather than an externally controlled orientation), those extroverted rather than intro-verted, those sensing rather than intuitive (preference for known facts versus possibilities), and those judging rather than perceiving (preference for flexible versus planned way of life) reported more positive attitudes toward teaching.

The single significant interaction identified in the ANOVA procedures suggested that the main effect of atti-tude development over the three points in career develop-ment cannot be interpreted accurately without considering the candidates’ locus of control orientation. The presence of the locus of control χ time in development interactions might in part explain some of the inconsistencies of find-ings in previous research. For example, in this study the low externality (internally controlled) candidates’ reported

Figure 1. Time in Teacher Development x Locus of Control Orientation


the most positive attitudes during preparation but the great-est reduction in positiveness of attitude between the end of teacher preparation and the end of the fifth year of teaching as compared to their mid or high externality cohorts. Conse-quently, it might be that some but not all teachers follow the developmental conception of teacher attitude. In other words, some teachers may report a more positive attitude toward teaching during their developmental years; whereas others report a less positive attitude during these years.

Greater diversity in attitude toward teaching was found at the end of the fifth year of teaching compared to the pre- and post-teacher preparation points in career development. For all classifications of the neophyte teachers, the standard deviations were greater for the fifth year of teaching atti-tude scores than for the attitude scores obtained prior to or following teacher preparation. The standard deviations for the attitude scores at the fifth year of teaching were approxi-mately twice the magnitude of the standard deviations for attitude scores obtained prior to or at the end of teacher prepa-ration within the various academic or personal classifica-tions of the teachers. This would appear to further support the suggestion that some teachers report more positive and some less positive attitudes toward teaching during their de-velopmental years.

In summation, attitude toward teaching as a career re-mained constant during teacher preparation but became less positive between the end of teacher preparation and near the end of the fifth year of classroom teaching for this sample of teachers. Perhaps this general decrease in positive attitude as well as the increased diversity of attitude scores noted at the fifth year of teaching reflects the challenge and demands of a profession characterized by stress, burnout, and high attrition rates as well as personal attributes including locus of control orientation. The attitudes of these candidates were found to be related to 6 of the 15 academic and personal characteristics selected for study. These findings support the findings of Tabachnick and Zeichner (Tabachnick and Zeichner, 1984; Zeichner, 1980) and further emphasize the role of personal traits in teacher development. Relatedly, and perhaps of particular concern, was the finding in the present study that the internally controlled candidates, noted in pre-vious research as feeling more responsible for pupils (Ashton, Webb, and Doda, 1983) and having higher achieving pupils (Murray and Staebler, 1974), reported relatively more abrupt declines in the positiveness of attitude toward teaching from the end of teacher preparation to the end of the fifth year of teaching than did their cohorts. This suggests that our more desirable teachers might most suffer in the transition from students to teachers and, as a consequence, may be most prone to leave the profession.

References

Ashton, P. T., Webb, R. B., and Doda, N. (1983). A study of teachers’ sense of efficacy. Final Report, Gainesville: Uni-versity of Florida, Contract No. 400-79-0025. National Institute of Education.

Bendixen-Noe, M. K. and Redick, S. S. (1995). Teacher de-velopment theory: A comparison between traditional-aged and nontraditional-aged beginning secondary teachers. Action in Teacher Education, 17(l), 52-59.

Berliner, D. C. (1988). Implications of studies of expertise in pedagogy for teacher education and evaluation. In New Directions for Teacher Assessment (Proceedings of the 1988 ETS Invitational Conference, pp. 39-68). Princeton, NJ: Educational Testing Service

Brookhart, S. M. and Freeman, D. J. (1992). Characteristics of entering teacher candidates. Review of Educational Research, 62(l), 37-60.

Bunting, C. (1988). Cooperating teachers and the changing views of teacher candidates. Journal of Teacher Educa-tion, March-April, 42-46.

Burden, R. R. (1982). Implications of teacher career devel-opment: New roles for teachers, administrators and pro-fessors. Action in Teacher Education, 4(3), 21-25

Byler, B. L. and Byler, L. F. (1984). Analysis of student teacher morale before and after student teaching. Jour-nal of the American Association of Teacher Education in Agriculture. 25, 22-28.

Callahan, R. (1980). A study of teacher candidates’ attitudes. College Student Journal, 14, 167-175

Chester, M. D. and Beaudin, B. G. (1996). Efficacy beliefs of newly hired teachers in urban schools. American Edu-cational Research Journal, 33, 233-257.

Firestone, W. A. and Pennell, J. R. (1993). Teacher commit-ment, working conditions, and differential incentive poli-cies. Review of Educational Research, 63(4), 489-525.

Fuller, F. and Bown, 0. H. (1975). Becoming a teacher. In K. Ryan (ed.), Teacher Education (Seventy-fourth Year-book of the National Society for the Study of Education, pp. 25-52). Chicago: University of Chicago Press.

Friesen, D., Prokop, C. M., and Sarros, J. C. (1988). Why teachers bum out. Educational Research Quarterly, 12(3), 10-19.

Gibson, R. L. and Coleman, D. R. (1997). Reviewing edu-cational reform and teacher education. Action in Teacher Education, 19, 84-87.

Hogben, D. and Lawson, M. J. (1984). Trainee and begin-ning teacher attitude stability and change: Four case stud-ies. Journal of Education for Teaching. 10(2), 135-153.

Hoy, W. K., and Rees, R. (1977). The bureaucratic social-ization of student teachers. Journal of Teacher Educa-tion, 28, 23-25.

Hoy, W. K. and Woolfolk, A. E. (1990). Socialization of student teachers. American Educational Research Jour-nal, 37, 99-118.

Hutchinson, G. E. and Johnson, B. (1993-94). Teaching as a career: Examining high school students’ perspectives. Ac-tion in Teacher Education, 15, 61-67.


Kagan, D. M. (1992). Professional growth among preservice and beginning teachers. Review of Educational Research, 62(2), 129-169.

Lipka, R. P. and Garlet, L. R. (1981). Age and intergroup differences in attitude toward the teaching profession: How do teachers and students view themselves and each other? Contemporary Educational Psychology, 6, 12-21.

Marso, R. N. and Pigge, F. L. (1997). A longitudinal study of persisting and nonpersisting teachers’ academic and personal characteristics. The Journal of Experimental Education, 65, 243-254.1

Marso, R. N. and Pigge, F. L. (1989). The influence of preservice training and teaching, experience upon atti-tude and concerns about teaching. Teaching and Teacher Education, 5, 33-41.

Murray, H. and Staebler, B. K. (1974). Teachers’ locus of control and student gains. Journal of School Psychol-ogy, 12, 298-310.

Myers, I. B. and McCaulley, M. H. (1985). Manual: A guide to the development and use of the Meyers-Briggs 1 Year Indicator. Consulting Psychologists Press, Palo Alto.

Noad, B. M. (1979). Influence of self-concept and educa-tional attitudes on elementary student teacher perfor-mance. Educational Research Quarterly, 4(l), 68-75.

Paschal, B. J. and Treloar, J. H. (1979). A longitudinal study of attitude change in prospective and beginning elemen-tary school teachers. Teacher Educator, 15(l), 2-9.

Pigge, F. L. and Marso, R. N. (1992). A longitudinal com-parison of the academic, affective, and personal charac-teristics of persisters and nonpersisters in teacher prepara-tion. The Journal of Experimental Education, 61, 19-26.

Pigge, F. L. and Marso, R. N. (1987). Relationships between student characteristics and changes in attitudes, concerns, anxieties, and confidence about teaching during teacher preparation. Journal of Educational Research, 81(2), 109-115.

Ramsay, W. and Ransley, W. (1986). A method of analysis for determining dimensions of teaching style. Teaching and Teacher Education, 2, 69-79.

Rotter, J. B. (1966). Generalized expectancies for internal versus external control of reinforcement. Psychological Monographs, 80, (Whole No. 609).

Sandgren, D. L. and Schmidt, L. G. (1956). Does practice teaching change attitudes toward teaching? Journal of Educational Research, 49, 673-680.

Sorenson, G., Schaefer, S., and Nyman, E. L. (1966). A teacher dropout for the MTAL California journal of Edu-cational Research, 17, 91-95.

Tabachnick, R. and Zeichner, K. (1984). The impact of the student teaching experience on the development of teacher perspectives. Journal of Teacher Education, 35(6), 28-36.

Wilkinson, G. A. (1994). Support for teacher induction. Action in Teacher Education, 16, 52-61.

Valli, L. (1992). Beginning teacher problems: Areas of teacher education improvement. Action in Teacher Edu-cation, 14, 18-25.

Villeme, M. G. and Hall, B. (1980). The relation of teacher attitude to major, employment status, teaching level, and satisfaction with teaching for first-year teachers. Human-istic Education, 19, 85-90.

Zeichner, K. M. (1980). Myths and realities: Field-based experiences in preservice teacher education. Journal of Teacher Education, 31, 45-55.


Introduction

Along with persistent demands for rigorous academic standards, a recurring theme in current calls for educational reform is the need to involve local schools and districts in planning for school improvement. A decade after the re-lease of A Nation At Risk, Terrel H. Bell, former Education Secretary and co-author of the report, acknowledged that as a result of the ineffective top-down reform characteristics of the 1980s, “changes in decision-making authority have been sweeping the nation” (1993, p. 595). Bell is unaware of any major American school system that does not have a campaign underway to strengthen site-based management of schools. In fact, varying models of educational decen-tralization are “in vogue” among many Western nations (Caldwell, 1992; Winkler, 1993). Although the extent to which site-based management is implemented varies from district to district, it is apparent that the trend toward local governance of schools is redefining the roles and responsi-bilities of teachers, parents, administrators, and other mem-bers of the local school community.

Review of Relevant Literature

A fundamental element of site-based management mod-els is one or several strata of school-level governing bodies that are comprised of administrators, teachers, parents, and community members. These stakeholders are presumed to be the most knowledgeable individuals when it comes to identifying and establishing effective learning environments for students at a particular school (Carlos and Amsler, 1993; Marburger, 1990; Mojkowski and Fleming, 1988). Behind the thrust for school-based governance councils is the prin-ciple of participatory democracy. Theoretically, greater par-ticipation in decision-making helps build consensus for certain reforms, thus ensuring that changes reflect the judg-ments and expertise of those directly involved in the teach-ing-learning process (Carlos and Amsler, 1993; Kreps, 1986; Owens, 1987; Ramirez, Webb, and Guthrie, 1991). As schools adopt participatory decision-making models, stake-

holders in the school improvement process have been com-pelled to develop skills and techniques that promote effec-tive group communication, consensus building, and problem solving (Bailey, 1991; Gresso and Robertson, 1992; Schlechty, 1990).

Not surprising, the emphasis on educators as decision- makers has fostered a renewed interest in the concept of action research. First popularized by Kurt Lewin in the 1940s, action research has endeavored to create a space in the research process for individuals who have been tradi-tionally viewed as non-researchers (i.e., teachers and ad-ministrators). Examples of exploratory action research activity have been many and varied, and research collabora-tions between schools and universities world-wide have been documented throughout the United States and abroad (King and Lonnquist, 1992, 1994; Noffke, 1997). Widespread in-terest and study notwithstanding, the use of action research, as a school improvement model, has yet to become either well understood or well established in schools.

Purpose and Context of the Study

This study considers both the process and product of the school improvement endeavors in a mid-sized urban school district. Several years earlier, the district had adopted a site-based management system in which administrators and teachers assumed responsibility for the educational outcomes in their buildings. To facilitate this restructuring effort, build-ing leadership teams were established in each school. The initial progress of the building leadership teams was slow and inconsistent, so the school district teamed with a local university to provide training for the school improvement teams. The intent of the training was to help building lead-ership teams develop communication and action research skills that would be useful for the teams as they developed a written school improvement plan. This study describes the collaborative efforts of the school district and university to provide the action research training. Specifically, the pur-pose of the study was twofold: (1) to identify the barriers and facilitators of the use of action research skills as a method

Using Action Research to Facilitate School Improvement Susan N. Kushner Sharon D. Kruse

University of Akron

Abstract This study describes a joint project between a mid-sized urban school district and local university. Specifically, district and university personnel collaborated to provide action research training for building leadership teams. A multi-method approach was used to identify the barriers and facilitators of the use of action research training as a method for teachers to create school improvement. A content analysis of the action research plans that were a summative product of the training is provided. Findings include the importance of both technical research skills and attention to the provision of time as a resource for those involved with school improvement efforts.


for teachers to create school improvement, and (2) to con-duct a content analysis of the school improvement plans that were a product of the action research training.

Methods

Participants. Fourteen building leadership teams (3 high schools, 3 middle schools, and 8 elementary schools) par-ticipated in the action-research training. Each building team was comprised of the building principal or assistant principal, a parent (usually a PTA member), teachers, and other support personnel (primarily guidance counselors). One of the teachers on each team served as the official union representative. In addition to the 104 school-based partici-pants, several district administrators, the union association president, representatives of the PTA, and three university professors assumed various roles throughout the training.

Action Research Training. The training was initiated during a two-week summer workshop. A senior district ad-ministrator and university professor assumed responsibility for developing the training sessions, and they selected a num-ber of commercially available materials (ASCD, 1994; Calhoon,1994; Sagor, 1992). The two-week training was actually conducted by a collaborative team of district and university personnel. The essence of the training was a five- step action research process of problem formation, data col-lection, data analysis, reporting results, and action planning. In addition to this technical training participants received training in team building and communication skills.

Following the summer training, monthly follow-up ses-sions were held. The purpose of these sessions was to pro-vide additional support and address specific issues that teams encountered as they began to implement the action research process in their respective schools. Teams were afforded the opportunity to share ideas and experiences and provide sup-port to members of other teams throughout the district. Dur-ing this four-month time period, facilitators were available to work with individual schools. A total of 38 site visits were made to 10 of the participating schools. These 10 schools had requested some form of support from a facilitator (e.g., assist with data analysis, lead focus group, etc.). Several months into the training, each team was required to develop and submit a written action research plan. Presently, dis-trict and university personnel are continuing to work with individual teams and at least one more group session is planned.

Procedures

A multi-method approach was used to evaluate the pro-cess and product of the action research training. First, a somewhat traditional closed and open-ended questionnaire was administered at the completion of the two-week train-ing program. The purpose of the instrument was to assess the participants’ perceptions of their readiness to undertake action research initiatives at their respective schools. The

questionnaire was administered several months later to gauge changes in attitudes and perceptions.

Next, interview data were collected and field notes were developed after each follow-up session and throughout the subsequent months as the building teams began to imple-ment the action research steps. The field notes included ses-sion agendas, participant’s work to date, and short interview style quotes that were generated during large group discus-sion (Yin, 1984, Merrimam, 1990). Focus group interviews (Kruger, 1988) were used to develop the teams shared per-ceptions of the action research process and were compared to the individual perceptions gathered (Miles and Huberman, 1994). From this data set teachers’ and administrators’ per-ceptions of the barriers and facilitators of employing an ac-tion research process to the study of school improvement were generated.

Finally, a content analysis of the action research plans was conducted. A thematic analysis of the manifest compo-nents of the plans revealed the status of the school improve-ment endeavors across the teams and provided evidence for evaluating the efficacy of the training.

Results and Discussion

Initial Training. At the conclusion of the summer work-shop training, the members of the school improvement teams were excited and motivated about the endeavor that lay ahead. They were generally very confident about their abili-ties to implement action research in their schools, and they were quite satisfied with the personal and professional de-velopment that the training afforded. The participants ex-pected that the knowledge and skills they gained during the training would be useful both during the initial stages of implementing school-based management and in sustaining their efforts. Many of the participants commented enthusi-astically on the partnership between the district and univer-sity personnel. The joint endeavor was seen as a positive and worthwhile venture. Without hesitation, the component of the training that they valued most was the time to work cooperatively within and across schools teams.

Follow-up Training Sessions. The analysis of the quali-tative interview data that were generated during subsequent training sessions suggests that time was the major barrier to successful development and implementation of a school improvement plan. Lack of time to think, talk among team members, and prepare and analyze data collection tools were cited as significant barriers to the improvement effort and hampered both individual and team goals.

Additionally, the qualitative data provided some insight into issues of teacher and principal empowerment as a result of the action research effort.

Time. If there is any consensus in the school improve-ment literature on the single most important structural sup-port teachers could receive, it is that of time (Caldwell and Spinks, 1992; Hargreaves, 1994; Little, 1993; Raywid,


1993). The findings of this study are no different. How-ever, what our results suggest is that additional time to work on improvement efforts is necessary, and the use of the time spent must be considered. As others have already found, the provision of time to meet and plan may result in changes in terms of curriculum planning and classroom innovation, it cannot guarantee innovation (Louis and Miles, 1990). The same is true of action research efforts. Adequate time to meet and study may improve the chances that quality re-search efforts will result; however, time alone cannot ensure success.

One way to consider this issue is to consider time as it is experienced in schools. For the teachers in this study, and teachers elsewhere, the lived time of the classroom and school day is intense. Daily, many decisions must be made and often little ‘research’ goes into the decision-making pro-cess that results in daily classroom bound decisions. Thus, it is natural for teachers to fall back on these kinds of deci-sion-making processes when approaching an action research project as well. However, as we reviewed the data collec-tion tools and subsequent analysis of school based data we found that quick immediate “peeks” at the information col-lected is not sufficient to create shared understandings of either the problems to be studied or solutions to be explored.

Teacher Empowerment. While teachers reported that they felt “empowered” by the research tools the training pro-vided, school politics and structural concerns (time to meet and plan, shared understandings of school mission and val-ues) dominated the focus group interview sessions. Fur-thermore, the focus group interviews suggested that the pressures of teaching in an urban district exacerbated even minor analysis efforts. In particular, meetings scheduled to analyze data often became “emergency” sessions to trouble-shoot the daily concerns of managing an urban school popu-lation. However, there were many bright spots among the interview data as well. Team members reported feelings of efficacy in both the use of “research skills” to analyze school- based data and the development of the team approach as a method to collectively examine school issues. Focus group data reinforced the emotional, social and intellectual sup-port members had gained during the study’s progress.

Principal Empowerment. While principals may have experienced some relief at the prospect of sharing the bur-den of school reform and change efforts, they also did not reevaluate what the role changes meant in terms of their positions of school leadership. Being freed from “having to know all the answers” in and of itself doesn’t free an admin-istrator from the role of keeping the school moving toward its improvement goals. Thus, they found themselves trapped in a common dilemma—how to empower staff while still leading the group toward shared reflection and progress. Moreover, this is a second example of a lack of technical skill albeit of a different kind. The principals lacked the technical skills to renegotiate their new roles in addition to the technical skills to prepare the action research report docu-mentation. Therefore, while they may have delighted in the

ability to “share the burden” they were unable to take the freedoms a lightened burden afforded and channel that en-ergy elsewhere in the school organization.

Analysis of the Action Research Plans. At the conclu-sion of the follow-up session, the teams were required to submit a written action research plan. It was made clear by the facilitators that the written plans were “working docu-ments” and that in all likelihood, teams would be at differ-ent points in the action research process. Although the teams were instructed to follow the guidelines suggested in the ASCD training materials, there was no attempt to prescribe the exact format that the teams should follow in writing their plans. It seemed reasonable to expect, however, that each plan would contain a problem statement, data collection and analysis methods, and action steps since both the ASCD materials and the facilitators’ directions explicitly identified these components.

Organization and Format. Although the plans did con-tain similar components, they were also characterized by a considerable amount of variability. Each plan was uniquely organized, varied in both the depth and breadth of informa-tion provided and the amount and type of ancillary materi-als included (e.g., mission statements, sample questionnaires, data analyses, etc.). An initial examination of the plans re-vealed two distinct organizational patterns. One group of plans (a total of eight plans) were logically-organized and professional-appearing of 10-15 pages in length. These build-ing teams documented their work with a variety of support materials, and all but one of the plans contained a list of the team members. Two of the plans in this group were submit-ted in nicely-bound notebooks with colorful covers and seemed to convey a special sense of school pride. In con-trast, the other group of plans (a total of six plans) clearly stood apart from the others. Of this group of six plans, five of the plans were more like executive summaries of the teams’ progress to date, and their lack of description docu-mentation (1-5 pages) or organizational structure made them difficult to analyze and evaluate. The sixth plan was simply a conglomeration of documents (i.e., minutes from meet-ings, memos, several questionnaires, a ballot of some sort, etc.). This plan had no logical order or structure to it.

Problem Statement. Each of the plans identified a prob-lem that would serve as the focus for the research plan. Only two of the plans described problems that were directly re-lated to learning “improving math proficiency” and “improv-ing language arts and writing” respectively. Interestingly, six of the plans identified achievement criteria as the goal (i.e., raising standardized test scores, increasing grade point average, and increasing attendance rates). Three schools identified goals that are somewhat tangential to student learn-ing (i.e., safety, and school climate). Finally, three teams identified the need to improve faculty communication as their problem. The focus on performance criteria such as test scores and attendance rates is noteworthy. One might argue that students who attend school regularly and perform well on standardized measures are more likely to have mastered


the skills and knowledge that will prepare them for the workforce or post-secondary education. Or, one might con-tend that schools today are pressed to use hard data to dem-onstrate “evidence” of school improvement.

Both the workshop facilitators and the training materi-als stressed the need to consider the professional literature as teams began to clarify and focus their problem statements. Nevertheless, only one team indicated that they planned to consider the literature related to their problem area. Of course, the absence of a reference to related literature can-not necessarily be construed to mean that teams omitted this process. Rather, it might simply be a conscious choice of the team not to document this step. Given that educators are often criticized for dismissing the value of professional lit-erature, however, it is just as likely that this all-important step may have been omitted.

Although the teams were able to identify a problem area for study, most problem statements were written in broad, generalized descriptions and only four teams included ac-tual research questions in their problem statement C even though the samples provided in the training materials clearly identified the need for research questions. While the lack of clearly identified research objectives could be an indication that the teams were still in the process of delimiting their problem area, it is more likely that the teams lacked the skills and practice to focus their target area—a task that is chal-lenging to even experienced researchers. The need to iden-tify specific questions, however, is a critical initial step in the action research process because it is the questions them-selves that will drive the data collection and analysis steps.

Sources of Data. The second component of the plans was sources of data, and 13 of the teams addressed this topic. The four teams that listed specific research questions were able to identify data that were congruent with their ques-tions. For example, the number of office referrals and inter-view data were suggested data sources as indicators of improved safety. Writing samples, test scores, and teacher observations were identified as indicators of written expres-sion. Nine teams did not identify specific research questions, and they took one of two approaches to this section. Six of these teams identified the data sources (e.g., staff, student, and parent questionnaires, archival data, and interviews, etc.) that they used to identify and clarify their problem focus. The remaining three teams conjectured about the sources of data that they might use as they implemented their plans.

Several issues bear noting. First, teams did seem to un-derstand that the value of multiple measures. Of particular concern however, is the obvious misuse of survey methods. Twelve teams proposed or actually sent questionnaires to parents. Judging by the number of responses reported, it is clear that there was no consideration given to the necessary sample size, the representativeness of the sample, and the biased responses.

Data Analysis. The five teams that collected data to clarify their problem statements were actually in a position

to analyze data. Indeed, three of these teams reported ac-tual results (in the form of frequency counts and narrative analysis) and the other two teams simply stated that they analyzed the results. The remaining teams could only con-jecture about how they might analyze the data, and they did so in very vague and broad terms.

Action Plan. The last component of the documents was an action plan. The purpose of this part was to identify the steps that teams expect to initiate. Although ten teams ad-dressed the topic of an action plan, only three of the teams actually identified a list of specific steps that they plan to undertake, and only one of the teams actually considered such topics as resources, participants, and a time-line for completing the tasks. The remaining seven plans included a very vague and often flowery narrative about future actions.

In summary, the variability in the plans may be due to several factors. First, given the nature of the training materi-als, it may be that the technical skills of action research were not sufficiently addressed. The need for choice and indepen-dence in developing their plans notwithstanding, beginning researchers are likely to benefit from structured guidelines or perhaps even a template to guide their efforts. Certainly, the technical skills of action research must be addressed.. To the trained researcher, the ASCD materials may make sense. But the process of systematic inquiry needs more depth. It is dif-ficult to adapt formal research methods (i.e., survey sampling, questionnaire design, etc.) for use in action research endeav-ors if one doesn’t know the fundamentals. Second, it was obvious through the interview and focus group data that teams were at varying stages. Some teams were cohesive units. Other teams appeared to struggle through the process. Clearly the need to communicate among themselves was prerequisite to any systematic problem solving.

Conclusion

In summary, our discussion results in two arguments. The first centers on the need for technical skills to be devel-oped in all members of the school improvement team for progress to be made toward an action research agenda. Sec-ond, it is important to note that the members of these teams felt strongly empowered and efficacious toward their ability to work together on shared common school improvement goals. Even without a complete battery of technical skills, teachers can achieve strong affective results by simply par-ticipating in a program focused on school improvement goals.

Time is not only necessary to carry out change agendas but essential if innovations like action research are to be maintained. Schools cannot remain both static and excep-tional. An institutionalized ongoing self-renewal process is necessary for the maintenance of school effectiveness, and this, in turn, implies a need for considerable and regular blocks of time devoted to technical skill-based learning and the creation of school improvement knowledge. Teachers need opportunities to consider action research plans and data


within department or grade level gatherings and in the con-text of all-school efforts. Consequently, the use of time de-voted to action research must be understood in two ways. First, teachers must be provided the means to meet on a daily basis to address issues of concern to immediate work groups of faculty—departments, grade levels or teams. Second, provision must be made for cross connection among smaller work groups that emerge in the full faculty. It is only when school efforts are clearly described and focused upon that improvement based on action research projects will occur.

References

Association for Supervision and Curriculum Development (1994). Action research: Inquiry, reflection, and deci-sion making. Facilitator’s Guide. ASCD: Alexandria: VA.

Baily, W. J. (1991). School-site management applied. Lancaster, PA: Technomic Publishing Company.

Bell, T. H. (1993). Reflections one decade after A Nation at Risk. Phi Delta Kappan, 74, 592-597.

Caldwell, B. J. and Spinks, J. M. (1992). Leading the self- managed school. The Falmer Press: Education Policy Perspectives Series.

Calhoon, E. F. (1994). How to use action research in the self-renewing school. Association for Supervision and Curriculum Development: Alexandria: VA.

Carlos, L. and Amsler, M. (1993). Site-based management: An experiment in governance (Policy Briefs No. 20). San Francisco, CA: Far West Laboratory for Educational Research and Development. (ERIC Document Reproduc-tion Service No. ED 356 522)

Gresso, D. W. and Robertson, M. B. (1992). The principal as process consultant: Catalyst for change. NASSP Bul-letin, 76, 44-48.

Hargreaves, A. (1994). Changing Teachers, Changing Times. New York: Teachers College Press.

King, J. A. and Lonnquist, M. P. (1992). A review of writ-ing on action research: 1944–present. Center for Ap-plied Research and Educational Improvement, Univer-sity of Minnesota.

King, J. A. and Lonnquist, M. P. (1994). The future of col-laborative action research: Promises, problems and pros-pects. Center for Applied Research and Educational Improvement, University of Minnesota.

Kreps, G. L. (1986). Organizational communication: Theory and practice. New York: Longman.

Kruger, R. (1988). Focus groups: A practical guide for ap-plied research. Beverly Hills: Sage Publications.

Little, J. W. (1993). Teachers’ professional development in a climate of educational reform. NCREST Reprint Se-ries. National Center for Restructuring Education, Schools and Teaching. Teachers College: Columbia University.

Louis, K. S., and Miles, M. (1990). Improving the urban high school: What works and why. New York: Teacher’s College Press.

Marburger, C. (1990). The school site level: Involving par-ents in reform (pp. 82-91). In Samuel B. Bacharach (Ed.), Educational reform: Making sense of it all. Boston: Allyn and Bacon.

Merrimam, S. (1990) Case study research in education: A qualitative approach. San Francisco: Jossey-Bass.

Miles, M. B. and Huberman, A. M. (1994). Qualitative data analysis: An expanded sourcebook. Thousand Oaks, CA: Sage Publications.

Mojkowski, C. and Fleming, D. (1988). School-site man-agement: Concepts and approaches. Cranston, RI: Rhode Island Educational Leadership Academy. (ERIC Docu-ment Reproduction Service No. ED 307 660)

Noffke, S. E. (1997). Professional, personal and political dimensions of action research. In Review of Educational Research. American Educational Research Association.

Owens, R. G. (1987). Organizational behavior in educa-tion. Englewood Cliffs, NJ: Prentice-Hall, Inc.

Ramirez, R., Webb, F. R., and Guthrie, J. W. (1991). Site- based management: Restructuring decision-making for schools (pp. 169-184). In James R. Bliss, William A. Firestone, and Craig E. Richards (Eds.), Rethinking ef-fective schools: Research and practice. Englewood Cliffs, NJ: Prentice Hall.

Raywid, M. (1993) Finding time for collaboration. Edu-cational Leadership, 51, 30-34.

Sagor, R. (1992). How to conduct collaborative action re-search. Association for Supervision and Curriculum De-velopment: Alexandria: VA

Schlechty, P. C. (1990). Schools for the 21st century. San Francisco: Jossey-Bass Inc.

Winkler, D. R. (1993). Fiscal decentralization and account-ability in education: Experiences in four countries (pp. 102-134). In Jane Hannaway and Martin Carnoy (Eds.), Decentralization and school improvement: Can we ful-fill the promise? San Francisco: Jossey-Bass Publishers.

Yin, R. (1984). Case study research: Design and method. Beverly Hills: Sage Publications.

The Mid-Western Educational Research Association’s

Annual Meeting

October 25–28, 2000 Holiday Inn—Plaza, Chicago, Illinois

The 2000 Annual Meeting of Mid-Western Educational Research Association has planned an exciting program of invited speakers, focused workshops, and paper presentations intended to generate discus-sion concerning education and educational research as we begin making a difference in the 21st century. Please join us and . . .

Look for us on the World Wide Web!

http://tierlab.ilstu.edu/MWERA

This WWW site provides conference information, including registration information, hotel reserva-tions, information about invited speakers, and abstracts of accepted presentations, along with links the to many highlights the City of Chicago offers.

A 1995 winner of the Harold W. McGraw, Jr. Prize in Education, Mary Diez is former President of the American Association of Colleges for Teacher Education and a leader in teacher education reform in the United States. She is a member of the Interstate New Teacher Assessment and Support Consor-tium, a group sponsored by the Council of Chief State School Officers and charged with the design of prototype standards for initial teacher licensure. In her writing, she focuses on standards and assess-ment in both teacher education and K-12 school reform. With colleagues at Alverno, she recently edited Changing the Practice of Teacher Education: Standards and Assessment as a Lever for Change (AACTE: 1998), which reports on a number of institutions engaged in the process of reconceptualizing their teacher education programs.

Featured Speakers

Dr. Mary Diez Will Reform Based on Standards and Assessment

Make a Difference in the 21st Century?

After outlining the elements of reform based upon stan-dards and assessment, Mary Diez will explore critical prin-ciples necessary for making a difference in learner outcomes. She will also address serious misconceptions and/or misap-plications that have under cut the reform’s potential both in K-12 schools and in teacher preparation.

Special Highlights

Fireside “Chat” with Mary Diez Join Dr. Diez for wine, cheese, and conversation about stan-dards development, assessment and reform in both K-12 education and higher education in America. What is our role in the areas of research, as well as public and political activism as we enter the 21st century? What can we do to make a difference in the 21st century?

A Forum for Best Practices Special sessions scheduled for Friday and Saturday will focus on what practitioner-scholars are doing to make a difference both for K-12 and higher education students.

A Joseph Panzer Professor of Education and Dean of the School of Education at the University of Dayton, Dr. Lasley is the past president of the Project 30 Alliance and currently secretary treasurer. Project 30 was an initiative of the AACTE and CCAS (Council of Colleges of Arts and Sciences) funded by the Carneige Foundation to interface the interests of education and arts and science faulty. Dr. Lasley served as the editor for the Journal of Teacher Education from 1984-1990, has published extensively in professional journals, and has co-authored seven books including the 1997 publication of Strategies for Teaching in a Diverse Society and in 2000, Strategies for Effective Teaching. During the past two years he has been principal investigator for a voucher based research project funded by the Smith Richardson Foundation and for a major national science foundation grant on science and mathematics.

Dr. Thomas J. Lasley II

“Why Teacher Education Fails: How It Can Succeed”

In recent years, teacher education has received endless criticism by the media, the public, the business community, and politicians. Dr. Lasley will speak to what the critics claim to be true about teacher education and suggest how we can begin to use research to enhance practice. He will address the need for colleges and schools of education to take risks, explore new paradigms for teacher education, and document what is occurring in teacher education so that the results can inform future practice.

Graduate Students—Special Topics A series of informal sessions will be provided for the expressed purpose of developing knowledge and skills in special interest areas for graduate students and new faculty members. Topics may include publish or perish, grant writing, and finding and securing that first faculty position

Division Meetings with Invited Speakers All MWERA Divisions will feature an invited speaker as part of the annual Division meetings.

The 2000 MWERA Annual Meeting will be held Wednes-day, October 25 through Saturday, October 28, at the Holiday Inn Mart Plaza in Chicago, Illinois. The program will consist primarily of presentations, selected through a peer review process, by divisional program chairpersons. In ad-dition, there will be invited speakers and symposia, panel discussions, special sessions for graduate students and new faculty, a luncheon and other social events open to all at-tendees.

Proposals may be submitted either on paper or elec-tronically over the World Wide Web. All proposals submitted on paper must be sent to just one Division. The Division Chairs’ addresses are noted below. Proposals must follow the Guidelines for Submitting a Proposal in this booklet. Questions about a proposal or the meeting, whether sub-mitted on paper or electronically, should also be directed to the Program Chair:

Dr. Carmen R. Giebelhaus MWERA-2000 Program Chair 300 College Park University of Dayton Dayton, OH 45469-0525 Office: (937) 229-4511 e-mail: [email protected]

Electronic proposals must be submitted using the form available on the meeting Web site. Proposals e-mailed to the Division Chairs or Program Chair will not be processed. Further, each proposal should only be submitted once in one format, electronic or paper. Specific instructions for elec-tronic submission can be found at the meeting web site:

http://tierlab.ilstu.edu/MWERA Any educational professional may submit a proposal

for MWERA-2000, whether or not that person is currently a member of MWERA. All Annual Meeting presenters must be members in good standing of MWERA (non-members must join MWERA upon notification of proposal acceptance). To promote broader participation in the program no one per-son should appear as a presenter on more than three pro-posals.

All proposals, regardless of submission format (elec-tronic or paper), must be received by the designated Divi-sion Chair no later than the deadline of May 1, 2000. Each Division Chair will coordinate a number of volunteers in a system of blind (without author identification) review. Appro-priate criteria, depending on the format and type of schol-arly work being presented, have been developed and are used for the review process. These criteria include: (a) topic (originality, choice of problem, importance of issues); (b)

relevance of topic to the Division and MWERA membership; (c) contribution to research and education; (d) framework (theoretical/conceptual/practical, rationale, literature review, grounding); (e) analyses and interpretations (significance, implications, relationship of conclusions to findings, generalizability or usefulness); and (f) overall written pro-posal quality (clarity of writing, logic, and organization).

Papers presented at MWERA are expected to present original scholarship, conducted by the author(s), which has not been previously presented at any other meeting or pub-lished in any journal. Further, it is a violation of MWERA policy to promote commercially available products or ser-vices (except as Exhibits) which go beyond the limits of ap-propriate scholarly/scientific communication. Individuals who wish to display educationally related products or services are encouraged to contact Dr. Sharon McNeely, Assistant Program Chair for Exhibits, P. O. Box 34421, Chicago, Illi-nois 60634, (913) 794-2788.

All persons presenting at the 2000 Annual Meeting are expected to register for the full meeting. All sessions listed in the program will be open to any registered meeting par-ticipant; however, enrollment may be limited, and a small additional fee required, for some Workshop sessions. Tick-ets for the Friday luncheon and speaker are available to all pre-registrants. Ticket availability is not guaranteed for late and on-site registrants. Registration materials for the 2000 Annual Meeting will be published in the Mid-Western Edu-cational Researcher, on the Web site, and can be obtained by contacting the Program Chair.

Presenters whose papers have been accepted to a ses-sion with a Session Chair and/or Session Discussant are responsible for submitting a completed version of their con-ference paper to the Session Chair and Discussant no later than September 20, 2000. Papers not available to the Ses-sion Chair and Session Discussant may be dropped from the program. Presenters must also provide complete copies of their papers (or detailed handouts) to attendees at their sessions. Overhead projectors and screens will be provided by MWERA in most presentation rooms. Presenters need-ing additional A/V equipment are responsible for arranging such with the hotel at the presenter’s own additional expense.

MWERA reserves the right to reproduce and distribute summaries and abstracts of all accepted proposals, includ-ing making such works available in a printed Program Ab-stract, through the meeting’s World Wide Web site, and in press releases promoting the Annual Meeting and the orga-nization. As a condition of acceptance all authors of papers accepted to the 2000 Annual Meeting explicitly grant MWERA

General Information

the right to reproduce their work’s summary and/or abstract in these ways. Such limited distribution does not preclude any subsequent publication of the work by the author(s).

Authors of accepted proposals assume the ethical and professional responsibility to appear at the Annual Meeting and to participate in their presentation or assigned session. When circumstances preclude the author(s) from doing so, it is the responsibility of the author to arrange a suitable substitute and to notify the Program Chair in advance.

Divisions A - Administration and Leadership This division is concerned with research, theory, develop-ment, and the improvement of practice in the organiza-tion and administration of education. Paper proposals should be mailed to the Sr. Chair of Division A: Micheal Supley, P.O. Box 610, Kingsville, TX 78364

B - Curriculum Studies This division is concerned with curriculum and instruc-tional practice, theory, and research. Paper proposals should be mailed to the Sr. Chair of Division B: Nancy G. Saunders, 4301 W. Riverside, Muncie, ID 47304

C - Learning and Instruction This division is concerned with theory and research on human abilities, learning styles, individual differences, problem solving, and other cognitive factors. Paper pro-posals should be mailed to the Sr. Chair of Division C: Cindy Campbell Dept. of Technology, Research, & Assessment, Northern Illinois University, DeKalb, IL 60115

D - Measurement and Research Methodology This division is concerned with measurement, statistical methods, and research design applied to educational re-search. Paper proposals should be mailed to the Sr. Chair of Division D: Janet Sheehan-Holt , Dept. of Technol-ogy, Research, & Assessment, Northern Illinois Uni-versity, DeKalb, IL 60115

E - Counseling and Development This division is concerned with the understanding of hu-man development, special education, and the application and improvement of counseling theories, techniques, and training strategies. Paper proposals should be mailed to the Sr. Chair of Division E: Linda Bakken, Wichita State University, ACES, Box 123, Witchita, KS 67260

F - History and Philosophy This division is concerned with the findings and method-ologies of historical research n education. Paper propos-als should be mailed to the Sr. Chair of Division F: Louise Fleming, 313 Bixler Hall, Ashland University, Ashland, OH 44805

G - Social Context of Education This division is concerned with theory, practice, and re-search on social, moral, affective, and motivational char-acteristics and development, especially multi cultural perspectives. Paper proposals should be mailed to the Sr. Chair of Division G: Anne Stinson, UWW – Division of Curriculum and Instruction, 800 Main Street. Whitewater, WI 53190

H - School Evaluation and Program Development This division is concerned with research and evaluation to improve school practice, including program planning and implementation. Paper proposals should be mailed to the Sr. Chair of Division H: Isador Newman, Univer-sity of Akron, College of Education, Akron, OH 44325- 4208

I - Education in the Professions This division is concerned with educational practice, re-search, and evaluation in the professions (e.g., medicine, nursing, public health, business, law, and engineering). Paper proposals should be mailed to the Sr. Chair of Di-vision I: Joyce Miller, Chemistry Dept., Mt. Vernon Nazarene College, 800 Martinsburg Rd., Mt. Vernon, OH 43050

J - Postsecondary Education This division is concerned with a broad range of issues related to two-year, four-year, and graduate education. Paper proposals should be mailed to the Sr. Chair of Di-vision J: Rodney Greer, Horribin Hall, 1 University Circle, Macomb, IL 61455

K - Teaching and Teacher Education This division is concerned with theory, practice, and re-search related to teaching at all levels and in-service and pre-service teacher education, including field experience supervision and mentoring. Paper proposals should be mailed to the Sr. Chair of Division K: Jim Salzman, 2570 East Wallings Road, Broadview Heights, OH 44147

Important Dates Proposal Submission Deadline

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . May 1, 2000 Notification of Acceptance

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . July 15, 2000 Papers to Session Chairs/Discussants

. . . . . . . . . . . . . . . . . . . . . . . . September 20, 2000 Meeting Registration and Hotel Reservations

. . . . . . . . . . . . . . . . . . . . . . . . September 15, 2000 MWERA 2000 Annual Meeting

. . . . . . . . . . . . . . . . . . . . . . . . October 25-28, 2000

Session Format Descriptions Paper Presentation

Paper sessions are intended to allow presenters the opportunity to make short, relatively formal presentations in which they overview their papers to an audience. Three to five individual papers dealing with related topics are grouped into a single session running from 1.5 to 2 hours. The presenter(s) of each paper is(are) allowed approxi-mately 15 minutes to present the highlights of the paper. A single Session Discussant is allowed approximately 15 min-utes, following all papers, for comments and critical review. A Session Chair moderates the entire session. Presenters are expected to provide complete copies of their papers to all interested audience members.

Roundtable Discussion/Poster

Roundtable Discussion/Poster sessions are intended to provide opportunities for interested individuals to partici-pate in a dialogue with other interested individuals and the presenter(s) of the paper. Presenters are provided a small table around which interested individuals can meet to dis-cuss the paper. Presenters may elect to provide small, table- top poster-type displays, ancillary handouts, or other table-top A/V materials to augment their discussions. Inter-ested individuals are free to move into and out of these discussions/posters as they wish. Presenters are expected to make available complete copies of the paper on which the roundtable discussion/poster was focused.

Symposium

A symposium is intended to provide an opportunity for examination of specific problems or topics from a variety of perspectives. Symposium organizers are expected to iden-tify the topic or issue, identify and ensure the participation of individual speakers who will participate in the session, prepare any necessary materials for the symposium, and Chair the session. It is suggested, though not required, that the speakers or symposium organizer will provide interested individuals with one (or more) papers relevant to, reflective of, or drawn from the symposium.

Workshop

Workshops are intended to provide an extended pe-riod of time during which the workshop leader helps partici-pants develop or improve their ability to perform some process (e.g. how to provide clinical supervision, using the latest features of the Internet, or conduct an advanced sta-

tistical analysis). Organizers may request from 1.5 to 3 hours, and are responsible for providing all necessary materials for participants. Many workshops are scheduled for Wednes-day afternoon, although others may be scheduled through-out the conference. Organizers may, if they wish, receive an honorarium based upon the number of paid participants in their workshop and the fee schedule.

Alternative Session

The form, topics, and format of alternative sessions are limited only by the imagination and creativity of the orga-nizer. These options are intended to afford the most effec-tive method or approach to disseminating scholarly work of a variety of types. Proposals for alternative sessions will be evaluated on their appropriateness to the topic and audi-ence, their suitability to meet the limitations of time, space, and expense for MWERA, and the basic quality or value of the topic. The organization of alternative sessions is respon-sible for all major participants or speakers, developing and providing any necessary materials, and conducting or medi-ating the session. Because a variety of approaches may be proposed within this category, alternative session propos-als should include a brief rationale for the alternative being proposed.

Best Practices Forum

The “Best Practices” sessions are intended to provide opportunities for individuals or groups to present “best” or “promising” practices impacting both K-12 and higher edu-cation. Highlighting unique and innovative programs that have demonstrated promise for improving and enhancing educational practice. Presenters will be grouped by similar topics to facilitate discussion between and among the groups and audience. Presenters are expected to make available complete copies of the paper on which the “Best Practices” session focused.

Materials to be Submitted

The following materials list applies to proposals sub-mitted on paper. Separate guidelines exist for electronically submitted proposals (see the Web site for details).

Proposal Cover Sheet

Six (6) copies typewritten with all items completed. Ses-sion descriptors must be chosen from the list of descriptors provided (see table to the right).

Guidelines for Submitting a Proposal

Summary

Six (6) copies of a two to three page summary for use in judging the merits of the proposal. Summaries can be single- spaced, but must be typed on 8.5" x 11" paper in no smaller than 10-point type using 1" margins. All copies of the sum-mary should include the title of the proposed session in the upper left-hand corner of the first page. On three of the sum-maries only include the name of the presenter, with his or her complete mailing address, telephone and FAX, and e- mail, in the upper right hand corner of the first page. Pro-posals, which do not meet these criteria, may be refused by the Program Chair without review.

Summaries for Paper and Roundtable Discussion/ Poster proposals should explicitly address as many of the following as appropriate, preferably in this order: (1) Objec-tives, goals, or purposes; (2) Perspective(s) and/or theoreti-cal framework; (3) Methods and/or techniques (data source, instruments, procedures); (4) Results and conclusions; and (5) Educational and/or scientific importance of the work.

Summaries for Symposium, Workshop, and Alterna-tive Session and Best Practices Forum proposals should explicitly address as many of the following as appropriate, preferably in this order: [1] Descriptive title of the session; [2] Objective, goals and purposes of the session; [3] Impor-tance of the topic, issue, or problem; [4] Explanation of the basic format or structure of the session; [5] Listing of the Presenter and Co-Presenter(s), with an explanation of each person’s relevant background and role in the session; [6] Anticipated audience and kind of audience involvement.

Abstract

Three (3) copies of a 100 - 150 word narrative abstract. The abstracts of accepted papers will be published the MWERA 2000 Annual Meeting Abstracts book, and will be available on the World Wide Web site. Abstracts must be type-written, single-spaced, using a 12 point Arial or Times Roman font. Use clear, precise language, which can be understood by readers outside your discipline. In the upper left hand cor-ner of each abstract page type the title of the paper, and the name and institutional affiliations of each author.

Envelopes

Four (4) stamped, self-addressed, business size (#10) envelopes. These will be used to inform you of: (a) receipt of the proposal by the Program Chair; (b) the decision about your paper’s acceptance; (c) your scheduled session time, Session Chair, and Session Discussant, and; (d) meeting registration and hotel reservation information.

Session Descriptors Ability Grouping Educational Policy Performance Assessment Accountability Educational Reform Philosophy Accreditation Elementary Schools Physical Education Achievement Equating Planning Action Research Equity Politics Adaptive Testing Ethics Postsecondary Education Administration Ethnicity Principals Admissions Evaluation Private Education Adolescence Experimental Design Problem Solving Adult Education/Development Facilities Professional Development Affective Education Factor Analysis Program Evaluation Aging Faculty Development Psychometrics Anthropology Family/Home Education Qualitative Research Aptitude Finance Race Artificial Intelligence Gay/Lesbian Studies Reading Arts Education Gender Studies Research Methodology Asian Education Generalizability Theory Research Utilization Assessment Gifted Education Restructuring At-Risk Students Governance Retention Attitude High Schools Rural Education Attribution Hispanic Education School/Teacher Effectiveness Bilingual/Bicultural History Science Education Black Education Indian Education Self-Concept Business Education Indicators/Information Systems Social Class Career Development Individual Differences Social Context Case Studies Information Processing Social Processes/Development Certification/Licensure Instructional Design/Development Social Studies Education Child Development Instructional Practices Sociology Classroom Management Instructional Technology Special Education Classroom Research Intelligence Staff Development Clinical Education International Education/Studies Standard Setting Cognition Item Response Theory (IRT) Statistics Cognitive Processes/Develop Language Comprehension/Devel Stress/Coping Collaboration Language Processes Structural Modeling Community Colleges Law/Legal Student Behavior/Attitude Comparative Education Leadership Student Cognition Compensatory Education Learning Environments Student Knowledge Comprehension Learning Processes/Strategies Student Teaching Computer Applications Life-Span Development Studying Computerized Testing Literacy Supervision Computers and Learning Literature Survey Research Conceptual Change Mainstreaming Teacher Assessment Constructivism Mathematics Education Teacher Characteristics Continuing Education Measurement Teacher Cognition Cooperative Learning Media Teacher Education/Development Counseling Medical Education Teacher Knowledge Counselor Training/Supervision Memory Teacher Research Critical Theory Mentoring Teaching Context Critical Thinking Meta-Analysis Technology Cross-Cultural Studies Metacognition Testing Curriculum Middle Schools Test Theory/Development Data Analysis Military Education Textbooks Decision Making Minorities Tutoring Demography Moral Education/Development Urban Education Desegregation Motivation Validity/Reliability Differential Item Functioning Museum Education Vocabulary Dimensionality NAEP Vocational Education Dropouts Networking Women’s Issues Early Childhood Organization Theory/Change Work Economics of Education Peer Interaction/Friendship Writing


Several studies (discussed and cited below) have shown that the ability to learn depends to a large degree on the ability to read. Usually, students who demonstrate superior reading comprehension and have larger reading vocabular-ies, obtain better grades. This paper represents a continua-tion and extension of the existing research to the study of business administration. The paper describes the results of measuring the reading abilities of 235 business students at a regional university and relates the results to the students’ cumulative GPAs.

Considerable research relates reading skills to student performance. Wood (1982) reported positive correlations between the Nelson-Denny Reading Test scores for over 1,000 college freshmen and their course grades in general psychology, general sociology, introductory speech and gen-eral English writing. Wood (1988) found that “College grades appear to be best predicted by previous grades, either in high school or college (r = .3 to .6). The next best predictors of college grades tend to be scores on reading tests or gen-eral aptitude tests (r = .2 to .4) such as the Nelson-Denny Reading Test and the tests of the American College Testing Program (ACT) and Scholastic Aptitude Testing Program (SAT)” (p. 2). Wood continues, “Nelson-Denny Test scores and scores from ACT and SAT tests do function as reading tests and do predict the kind of college success that is mea-sured by college grades” (p. 5).

Kessler and Pezzetti (1990) studied the impact of read-ing ability on exam performance in six psychology classes. They report, “While the results varied for each instructor, read-ing ability was shown to be significantly related to test per-formance. On the average, the high reading group out-performed the low group by 7 to 12 percentage points on each major exam. While test scores improved over the se-mester for both groups, the low readers never bridged this performance gap. As students drop over the course of the year (for a myriad of reasons), the number of poor readers declines at a greater rate than those with better skills” (p. 1).

Similarly, Iadevaia (1989) reported the results of using a Pearson correlation coefficient calculation in a study of 558 students conducted between 1983 and 1988. He found a correlation of (r = .32539) between grades in an algebra- based physics course and the Nelson-Denny Reading Test results. He notes, “A possible conclusion was that the Nelson- Denny Reading score was the best predictor of a student’s completion of PHY 121 with a passing grade” (p. 17). Iadevaia’s results may be particularly interesting to business faculty trying to predict student success in quantitative busi-ness courses such as finance that depend heavily on the use of algebra.

Brown, et al. (1993) note, “The ability to read well is so important in our culture that it is often the principal cause of success or failure from the first grade of elementary school through college and into professional life” (p. 1). They con-tinue, “The University of Minnesota collected data over a five-year period. Test scores were collected for three groups—those graduating with high distinction, those with distinction, and those on probation. As incoming freshmen, those graduating later with high distinction averaged at the 83rd percentile on the Nelson-Denny test, those with dis-tinction at the 68th percentile, and those on probation at the 42nd percentile” (p. 8).

Existing research indicates that reading ability is im-portant to academic achievement. Students who demonstrate superior reading skills appear to have a learning advantage.

Research Design

During the end of the spring 1997 semester, a study was conducted to measure the reading ability of business stu-dents. The study included 235 participants, 131 females and 104 males, enrolled in day and evening classes. There were 100 sophomores, 86 juniors and 49 seniors. Students were tested using the Nelson-Denny Reading Test that measures reading comprehension, reading vocabulary, and reading

Reading Ability as a Predictor of Student Success in Business School

Robert E. Pritchard and George C. Romeo Rowan University

Abstract For several years accounting faculty at a regional university observed that some of their students exhibited difficulty reading. This study of 235 sophomore, junior, and senior business majors enrolled in accounting classes examines the relationship between their reading abilities as measured by the Nelson-Denny Reading Test and their cumulative grade point averages (GPAs). The results indicate that students who have higher levels of reading comprehension and reading vocabulary have higher cumulative GPAs. Furthermore, the results indicate that about 16 percent of the students tested could not read at the first-year college level. This paper describes the testing and recommends that students’ reading abilities be factored into the admission processes.


rate. The testing was conducted during regularly scheduled class periods by university faculty.

The Nelson-Denny Reading Test was chosen as the mea-surement tool primarily because of its demonstrated usefulness in predicting academic success. In addition, it was chosen be-cause of its wide acceptability over many years of use, its re-cent standardization on relevant populations (high school students and students from two-year and four-year colleges), and its minimization of gender and ethnic bias. In addition, the three stratifying variables (region, district enrollment, and so-cioeconomic status) were utilized in selecting participants in the standardization trials (Brown et al., 1993).

Research Analysis

The students’ scores using the Nelson-Denny Reading Test were analyzed using Chi-square, Pearson correlation coefficients and linear regression. Chi-square was used to test goodness of fit. The Pearson correlation coefficients were used to measure the strength of linear relationships. Linear regression was used to study the relationships among the variables as well as to summarize the data.

For Chi-square analysis, student scores on the reading comprehension, vocabulary, reading rate, and calculated reading grade equivalent (comprehension plus vocabulary) were divided into quintiles. The students’ cumulative GPAs were also divided into quintiles. Then, the cumulative GPAs were compared with outcomes in each of the reading-mea-sure quintiles. Continuous data were analyzed using the Pearson correlation coefficients and linear regression analy-sis. Relationships were considered to be statistically signifi-cant at the .05 level.

Research Results

The research indicates statistically significant relation-ships between the dependent variable (students’ cumulative GPAs) and the following three explanatory variables: 1) reading comprehension, 2) reading vocabulary and 3) cal-culated reading grade equivalent. However, a statistically significant relationship did not exist between students’ read-ing rates and their cumulative GPAs.

The descriptive statistics underlying the calculated read-ing grade equivalents for the 235 students who were tested are included in Table 1. The Pearson correlation coefficients are shown in Table 2. The regression analysis results are shown in Table 3. The relationships between the four ex-planatory variables (students’ reading comprehension, read-ing vocabulary, calculated reading grade equivalents, and reading rate) and their cumulative GPAs are shown in Tables 4 through 7 respectively.

As shown in Table 1, the students’ scores indicate a fairly wide dispersion of calculated reading grade equiva-lents, ranging from a low of 4.1 to a high of 18.9 (the high-est reading grade equivalent using the Nelson-Denny Reading Test), with a standard deviation of 3.32. Interest-

ingly, 11 students (nearly five percent) had calculated read-ing grade equivalents of 18.9 (the highest obtainable calcu-lated reading grade equivalent). Approximately 16 percent of the students in the study (38 students) had calculated read-ing grade equivalents below the thirteenth grade level.

The relationship between the students’ reading compre-hension scores and their cumulative GPAs was statistically sig-nificant at the .05 level using Chi-square (P > .00054) as well as when using Pearson correlation coefficients, (r = .2680).

Table 1 Students’ Calculated Reading Grade Equivalents

Descriptive statistics Descriptive statistic Calculated read grd. equiv.

Minimum 4.1 Maximum 18.9 Range 14.8 Mean 14.92 Standard Deviation 3.32 Median 15.6 Mode 16.3

Percentile Calculated read grd. equiv. 25th 13.8 50th 15.6 75th 17.3

Table 2 Correlations Between Students’ Cumulative GPAs and Vocabulary, Reading Comprehension, Reading Rate, and Calculated Reading Grade Equivalents

Pearson correlation coefficients and probabilities Voc. Compre. Rate Rd. Gd. Eq. Cum.GPA

Voc. -- r=.7228 r=.3350 r=.9434 r=.2878 p=.000 p=.000 p=.000 p=.000

Compre. -- r=.3182 r=.8810 r=.2680 p=.000 p=.000 p=.000

Rate -- r=.3741 r=.0871 p=.000 p=.218

Rd. gd. eq. -- r=.3076 p=.000

Cum. GPA -- Note: Table heading abbreviations: Voc. = Vocabulary, Compre. = Comprehension, Rd. gd. eq. = Reading grade equivalent, Cum. GPA = Cumulative GPA.

Table 3 Cross-sectional Regression Results with Student Cumulative GPAs as Dependent Variable and Reading Vocabulary, Reading Comprehension, and Calculated Reading Rate Test Scores as Independent Variables GPA = b0 + b1 Vocab. + b2 Comp. + b3 Rate

Bvalue 2.430 .007 .005 9.537e-:05 t-ratio 14.363 2.204 1.464 -.208 Prob. .0000* .0285* .1445 .8351 * Significant at the .05 level

R2 F-Value Prob. .09150 7.55389 .0001


The relationship between the students’ reading vocabu-lary test scores and their cumulative GPAs was statistically sig-nificant at the .05 level using Chi-square (P > .00037) as well as when using Pearson correlation coefficients, (r = .2878). The regression analysis (Table 3) indicates that reading vocabulary is the most important factor affecting cumulative GPAs.

As noted above, calculated reading grade equivalent rep-resents a combined measure of reading comprehension and reading vocabulary. Since the students’ cumulative GPAs were very highly correlated with their reading comprehen-sion and reading vocabulary scores, it is not surprising that their cumulative GPAs are also highly correlated with their calculated reading grade equivalents at the .05 level using Chi-square (P > .00000) as well as when using Pearson cor-relation coefficients, (r = .3076). Clearly, those students who demonstrate superior reading comprehension and have larger reading vocabularies have higher cumulative GPAs.

The relationship between the students’ reading rate scores and their cumulative GPAs was not statistically sig-nificant at the .05 level using Chi-square (P > .54073). Nor was it significant when using Pearson correlation coefficients. The data are presented in Table 7.

Conclusions And Recommendations

The study results clearly indicate statistically signifi-cant relationships between students’ reading comprehension and reading vocabulary scores on the Nelson-Denny Read-ing Test and their cumulative GPAs. Students who have higher levels of reading comprehension and larger reading vocabularies generally have higher cumulative GPAs. Fur-

thermore, students’ cumulative GPAs were highly correlated with their calculated reading grade equivalents at the .05 level with P > .00000. However, although the relationship between reading ability and students’ GPAs is statistically significant as shown in Table 3, reading ability only explains about nine percent of the variation in students’ GPAs. Obvi-ously, there are other factors that affect student performance. As noted earlier, Wood (1988) indicated that college stu-dents’ grades are best predicted by previous high school and college grades.

The students’ reading rate scores were not statistically significant predictors of their cumulative GPAs. This is not surprising since understanding highly technical materials may take considerable time, including, for example, performing the calculations associated with accounting or finance text-book examples. Materials written for different purposes should be read at different rates. Thirty-nine percent of the students with reading rates in the two lowest quintiles had cumulative GPAs in the highest quintile. Reading compre-hension and vocabulary are the clear predictors of students’ cumulative GPAs, not reading rate.

Many colleges already test entry-level freshmen for reading skills. Weiner and Bazerman (1995) report that when information about student reading level is sought, it is com-monly taken from basic skills reading tests required by the state for entering college freshmen. Examples of these in-clude the Texas Academic Skills Program, the Georgia Col-legiate Placement Examination, and the Florida College Level Academic Skills Test. These tests are screening de-vices used to place students in regular or remedial tracks. Unfortunately, the existing admissions testing processes (and

Table 4 Students’ Cumulative GPAs and Reading Comprehension Test Scores

Reading comprehension test scores (quintiles) Lowest Highest

GPAs (quin.) 1 2 3 4 5 Lowest 1 24.4% 48.6% 18.5% 10.0% 2.9%

2 31.1% 8.6% 22.2% 15.0% 22.9% 3 24.4% 14.3% 16.7% 21.7% 22.9% 4 8.9% 17.1% 22.2% 26.7% 22.9%

Highest 5 11.1% 11.4% 20.4% 26.7% 28.6% Note: Table heading abbreviation: quin. = quintiles.

Table 5 Students’ Cumulative GPAs and Reading Vocabulary Test Scores

Reading vocabulary test scores (quintiles) Lowest Highest

GPAs (quin.) 1 2 3 4 5 Lowest 1 28.3% 34.9% 17.6% 15.2% 2.3%

2 26.1% 20.9% 11.8% 21.7% 20.9% 3 26.1% 27.9% 17.6% 17.4% 11.6% 4 10.9% 7.0% 29.4% 23.9% 27.9%


Table 6 Students’ Cumulative GPAs and Calculated Reading Grade Equivalents

Calculated reading grade equivalents (quintiles) Lowest Highest

GPAs (quin.) 1 2 3 4 5 Lowest 1 24.4% 47.7% 11.3% 10.8% 6.0%

2 31.1% 20.5% 15.1% 10.8% 22.0% 3 26.7% 11.4% 24.5% 27.0% 12.0% 4 8.9% 9.1% 24.5% 29.7% 28.0%


Table 7 Students’ Cumulative GPAs and Reading Rates

Reading rates (quintiles) Lowest Highest

GPAs (quin.) 1 2 3 4 5 Lowest 1 23.4% 24.1% 16.1% 15.4% 17.8%

2 23.4% 14.8% 25.8% 25.0% 13.3% 3 21.3% 22.2% 12.9% 21.2% 20.0% 4 17.0% 14.8% 35.5% 15.4% 24.4%



remediation programs) do not assure that all students can read at the college level.

One reason for this problem may be that the basic skills reading tests are typically not power tests that show deeper levels of vocabulary and inferential reading ability. Unlike many basic skills reading tests, the Nelson-Denny Test scores, for example, indicate those students who need fur-ther reading vocabulary development as well as those who have difficulties with reading comprehension. Consequently, the use of a test like the Nelson-Denny Reading Test as a part of a business admissions process could both identify more accurately those students who are poor readers as well as substitute for basic skills reading tests.

The high degree of correlation between reading grade equivalents as an explanatory variable and students’ cumu-lative GPAs, as well as the very broad range of student cal-culated reading grade equivalents (4.1 to 18.9), further suggests that students’ reading grade equivalents be mea-sured as a part of an additional screening process for admit-ting students to business programs (and perhaps other programs as well). Students who cannot read at the thirteenth grade level (or some other minimum level) could be required, for example, to take remedial courses in reading and/or ac-cepted for admission on a probationary basis.

In summary, the data point to the critical importance of students having good reading skills if they are to succeed in business programs. Furthermore, this study shows that ba-sic skills screening does not identify or eliminate all mar-ginal or needy students. Consequently, some students with weak reading skills are admitted to business programs. An important issue is how to assure they receive the help they need to succeed.

Brozo (1990) notes that reading is an interactive pro-cess. “This means that a student’s ability to comprehend is not fixed or constant; rather, comprehension will vary across texts, tasks and settings (prior knowledge and interest are two powerful factors that contribute to variability on read-ing tasks)” (p. 523). Brozo further states, “The goal of inter-active (reading) assessment, therefore, is to discover the conditions under which a student will succeed in reading, rather than merely describing a student’s current status as a reader” (p. 523). “The goal of assessment is not the identifi-cation of a disability but rather the specification of the con-ditions under which a particular student will learn” (p. 527).

Brozo (1990) notes, “Many poor secondary readers bring to the classroom a long history of failure and, likely, a repertoire of strategies designed to avoid reading, so solu-tions are not simple” (p. 327). The article includes a series of five strategies that can be employed to improve students’ reading and learning that may be applicable at the college level as well as at the secondary level.

Finally, it is important to note that there are differences in the performance on reading tests among African-Ameri-cans, Latino, and Native American students, as well as stu-dents for whom English is a second language. The critical factors underlying these differences are poverty and English proficiency (Garcia and Pearson, 1992). They note, “Dif-ferences between Anglo students and students of color are substantially reduced when comparisons are limited to stu-dents from the same income levels and similar proficiency in standard English” (p. 340). The development and imple-mentation of policies to measure business students’ reading ability and to provide programs to assist them need to con-sider these differences.

References

Brown, J. I., Fishco, V. V., and Hanna, G. (1993). Nelson- Denny Reading Test, manual for scoring and interpreta-tion. Chicago: The Riverside Publishing Company.

Brozo, W. G. (1990, February). Hiding out in secondary content classrooms: Coping strategies of unsuccessful readers. Journal of Reading.

Brozo, W. G. (1990, April). Learning how at-risk readers learn best: A case for interactive assessment. Journal of Reading.

Garcia, E. and Pearson, P. D. (1992). Assessment and diver-sity. In L. Darling-Hammond (Ed), Review of research in education (pp. 20-38). Washington, D. C.: American Education Research Association.

Iadevaia, D. G. (1989, February). A study of the relation-ship between student placement test scores and final grades in Physics 121 at Pima College. A practicum pre-sented at Nova University in partial fulfillment of the re-quirements for the degree of Doctor of Education.

Kessler, R. P. and Pezzetti, G. (1990, August). The impact of reading ability on classroom performance. (Research, Planning and Resource Development Report). Santa Ana, CA: Rancho Santiago Community College.

Weiner, H.S. and Bazerman, C. (1985). All of us: A multicultural reading skills handbook. Princeton, NJ: Houghton Mifflin.

Wood, P. H. (1982). The Nelson-Denny Reading Test as a predictor of college freshmen grades. Educational and Psychology Measurement, 42, 575-583.

Wood, P. H. (1988, September 17). Predicting college grades and helping colleagues to assist poor readers to succeed in college courses. Part of a two-paper presentation to the National Reading and Language Arts Educators’ Conference. Kansas City, MO.


In the last two decades major structural changes in the larger world environment have required substantial effort on the part of all organizations to survive, remain viable, and demonstrate their strategic effectiveness in the face of increasing uncertainty and public scrutiny. Vast external environmental changes in technology, demographics and economic structures have been juxtaposed with the increas-ingly tenuous nature of available resources necessary for successful professional education (Axinn and Thach, 1996). The overall effectiveness of one professional program ver-sus another is heatedly debated in published rankings, gradu-ate salaries and job offers with attendant disputes over academic accreditation. Part of the problem with measur-ing the effectiveness of professional programs is their dual role as academic and professional institutions serving two different constituencies, one academic and the other in prac-tice. Service to the academic discipline (e.g., rigor in theory and scholarship) competes with service to practice (e.g., applied problems and operations) (Cheit, 1985). As new societal demands are place upon professional programs, it becomes imperative that systematic analysis is extended not simply to the larger higher education institution in which the professional program resides but to the professional pro-gram itself.

Establishing measurements of effectiveness for profes-sional programs in general is of little use as a “diagnosis” for determining effectiveness in individual institutions without an understanding of why the school exists in such a state. Two central concepts thought to be necessary to an understanding of organizational effectiveness are a) that Management must make strategic choices that are fundamental in the determina-tion of organization design and serve to define the organization’s relationship with the broader environment (Child, 1972) and b) that organizational actions are guided and determined by the internal organizational culture that re-lates to the specific behaviors regarding decision making and acceptance (Katz and Kahn, 1978). This paper examines the

relationship of organizational culture and managerial strat-egy to each other and to organizational effectiveness.

Outcome Variables

Organizational Effectiveness. Researchers have devel-oped a number of multidimensional measures of effective-ness to describe the process by which decision makers determine the overall performance of the organization to its external and internal environments (Cameron, 1978; Peters and Waterman, 1982; Quinn and Rohrbaugh, 1983). Virtu-ally all studies and theories about organizations are in some way based on the construct of effectiveness despite the con-fusion surrounding its meaning. Cameron’s (1978) construct of organizational effectiveness is a theoretical framework that incorporates the use of multiple perspectives that can be recorded in relationship to one another and is the per-spective used in this study.

Organizational Culture. Organizational culture has been found to be a central concept in previous research studies seeking to understand and improve organizational develop-ment (Denison and Spreitzer, 1991, Ouchi 1981; Deal and Kennedy 1982; Peters and Waterman 1982; Schein 1985, 1990). Pettigrew (1979) defines organizational culture as the “amalgam of beliefs, ideology, language, ritual and myth” (1979:571) influencing internal organizational constituen-cies. Without an understanding and comprehension of the effects of organizational culture a true measure of organiza-tional performance is not possible (Wilkins and Ouchi 1983). Schein (1985) suggested that culture contributes to external adaptation of the organization through providing a consen-sus about mission, strategy and goals and to internal inte-gration issues through consensus of language, conceptual categories and group boundaries, influence and power. In essence, culture shapes action by defining what people want (Parsons, 1951; Swidler, 1986).

Many definitions of culture have been offered to ex-plain its meaning in organizations. Cameron and Ettington

Culture and Strategy in Business Schools: Links to Organizational Effectiveness

Christopher Clott Elmhurst College

Nancy Fjortoft Midwestern University

Abstract The purpose of this study is to examine the independent and conditional effects of organizational culture type and managerial strategy on the organizational effectiveness of higher education schools of busi-ness. While significant differences are found for both culture type and managerial strategy on the effectiveness measures, there were no significant interactive effects between the two independent vari-ables. The implications of these findings for efforts to enhance the effectiveness of professional and pre- professional collegiate programs are discussed.


(1989) found at least 17 definitions to suggest the variety and lack of consensus on meaning. Research by Ouchi (1980, 1981, 1983) suggested three types of organizational cultures grounded in transaction cost theory 1) Clans—Socializa-tion of all organizational members to a common goal with a close interchange between work and social life. 2) Mar-kets—Measuring contributions of organizational members to the common good with compensation based upon contri-bution and competitive bidding for services and 3) Hierar-chies—Contractual specifications of rights and duties to organizational members but with a lack of a clear sense of the value of their services. Ouchi argued that Market cul-tures succeed when numerous buyers and sellers succeed in maximizing their value through specific short term contrac-tual relationships e.g., performance of a specific project usu-ally in a given period of time. Market cultures fail under complex and large transactions and where the determina-tion of value contributed by an individual worker is ambigu-ous in the making of the final product. Hierarchies (bureaucracies) work well when long-term relationships, division of specialized labor and monitoring of employees is necessary to provide skills and expertise for complex out-comes. Hierarchies fail to the extent that a hierarchical sys-tem of “surveillance, evaluation and direction” fails to provide reasonable performance information to employees creating mistrust between employee and employer and when tasks become highly unique and therefore impossible to evaluate in any meaningful way. Clan cultures succeed where there is a high degree of internal cohesion and goal congru-ence among organizational members as well as a high de-gree of discipline and overlap of individual and organizational interests. Clan cultures fail when subjected to the scrutiny of contractual relationships where individual performance is specifically negotiated (Ouchi, 1980; 1981).

Researchers have also suggested that strong, externally oriented cultures are more effective than weak, internally oriented cultures (Cameron, 1987; Arnold and Capella, 1985). Cameron and Ettington (1988), Zammuto and Krakower (1991), and others have suggested for example, the existence of “strong” cultural types permeated by as-sumptions of achievement, growth, and resource acquisition that facilitate rather than impede strategy. An alignment between the beliefs and values central to the organization and actual strategies is the central distinguishing feature of strong cultures (Dennison, 1990).

Managerial Strategy. The concept of “strategy” is broad and overarching, encompassing numerous older and newer conceptualizations including strategic planning (Chandler, 1962), strategic choice (Child, 1972), strategic management (Ansoff and Hayes, 1976), and strategic norms (Hatten, 1979). In a review of various concepts of strategy, Chaffee (1986) suggested the existence of three distinct models in the literature. Linear strategy, the oldest concept, involves setting long term goals through the use of planning and fore-casting. Resources are then allocated to achieve organiza-tional goals, which are primarily profit and productivity.

Adaptive strategy is concerned with developing a viable match between risks and opportunities in the dynamic ex-ternal environment consisting of events, trends, competitors and organizational stakeholders. Adaptive strategy is more complex than linear strategy and considers more variables. It is by far the most heavily researched. The third model suggested by Chaffee is interpretive strategy, the newest concept. Interpretive strategy suggests individuals with free will enter into the organization. The organization’s contin-ued existence lies in its legitimacy and ability to attract and keep individuals through symbols, interactions and relation-ships. Each model provides a means of distinguishing cer-tain aspects of organizational functioning in which the concept of strategy has been applied. Chaffee (1986) has also suggested that a broad view of all research dealing with an aspect of strategy should be taken rather than attempting to select a narrow definitional segment that may exclude relevant research with slightly different definitions.

One typology of strategy measures implied in previous research on organizations is the existence of active and pas-sive types of strategy (Miles and Snow, 1978; Snow and Hrebiniak, 1980; Cameron, 1986; Clott, 1994). Typically, passive organizations are reactive in nature while active or-ganizations are constantly stimulating organizational change. An active strategy seeks to energetically manipulate the en-vironment and construct it to suit the needs of the organiza-tion. Overall, active strategies have been more important in mitigating the negative effects of the turbulent environments now found in colleges and universities (Cameron and Tschirhart, 1992). Cameron’s (1986) study of administra-tors in colleges and universities suggests that when faced with a depletion of resources there is a tendency for organi-zations to be become conservative and internally oriented, overly focused on cost containment, and reactive. The pas-sive organization has been hypothesized by Snow and Hrebiniak (1980) to be successful only for short periods of time within stable, protected and relatively benign environ-ments and “...generally not viable in competitive industries” (p.333). A further description of the passive strategy type suggests that it narrowly focuses on safeguarding the orga-nizational niche at all costs (Hambrick, 1983). The linkage between managerial strategy and organizational effective-ness has also been noted in studies by Cameron and Zammuto (1983), Miles and Cameron (1982) and Doty, Glick and Huber (1993) as most strongly associated with improving effectiveness over time. The strategic emphases and choices of managers are considered by researchers to be critical fac-tors in assessments of organizational effectiveness.

Culture and Strategy. The relationship between organi-zational culture and strategy has been addressed extensively in the literature (Ouchi, 1981; Deal and Kennedy, 1982; Tichy, 1983, Dennison, 1991). There has been little research in the organizational literature though, as to the exact cau-sality of the relationship between the two variables. Much research is grounded for example, in the assumption that cultural values and patterns of behavior can limit and con-


strain the implementation of managerial strategy. Tichy (1983) and Zammuto and Krakower (1991) have addressed the implementation process as the need to align the overrid-ing culture to strategic shifts in organizations. Davis and Schwarz (1981) suggested that a corporation’s culture fil-ters top management perspectives limiting the strategic op-tions they are prepared to consider seriously. They further argued that defining the cultural values of a company cul-ture can remove old constraints on past strategic decision making.

While the weight of the evidence appears to suggest that culture precedes strategy, the implication in general or-ganizational theory that culture exists merely as a limiting factor to strategy does not take into account the existence of varying types of cultures. Research generally has been in-consistent and difficult to interpret in this area due to varied and often confusing empirical studies of culture and strate-gic types. Therefore, this study examines the independent and conditional effects of organizational culture and mana-gerial strategy on the organizational effectiveness of a sample of business schools.

Methods. The subjects chosen for study were the aca-demic deans of schools of business and schools of accoun-tancy in the United States and Canada. Their titles included dean, director, or chairperson. These subjects were chosen because they are presumed to be the primary individuals involved with the external and internal environment of the organization and are considered to represent the “dominant coalition” in schools of business. Dominant coalition mem-bers are the key persons to determine managerial strategy as it impacts on the institution (Yuchtman and Seashore, 1967; Pennings and Goodman, 1977). A survey sample of 806 deans of graduate and undergraduate schools of business and accountancy comprising the entire population of Ameri-can Assembly of Collegiate Schools of Business (AACSB) member institutions for the United States and Canada was selected for the study. Three hundred thirty completed sur-veys were returned for a 41 percent response rate. There was one response per school. Approximately 64 % of the respondents were from public institutions and 36% from private. Respondents were from small, 30.5% (50-1000 stu-dents enrolled in undergraduate/graduate business programs), medium, 38% (1001-2500 students) and large, 31.5% (over 2500 students) institutions. Approximately 21% of the re-spondents were deans of schools of business with only bach-elors degree programs while 57% were deans of schools offering both bachelors and masters degree programs. Twenty-one percent (21%) were deans of schools with un-dergraduate, masters and doctoral degree programs. Demo-graphic data on the respondent characteristics suggest a diversity of age, experience and work activity. Data was not collected on the number of faculty per college.

Variables. Scales reflecting organizational culture, stra-tegic orientation and seven dimensions of organizational effectiveness were developed from items on the survey based

on previous research in effectiveness by Cameron (1986), Zammuto and Krakower (1991), and Cameron and Tschirhart (1992). The measure of organizational culture type of school of business in this study was derived from research performed by Ouchi (1980, 1981). The three types of culture utilized were 1) A Clan culture type identified through an emphasis on shared goals and values through consensus decision making. 2) A Hierarchy type identi-fied through an emphasis on order, authority and efficiency. 3) A Market type identified through an emphasis on com-petitiveness, goal accomplishment and efficiency and through the mechanism of the marketplace. The three cul-ture types and coefficient alphas for their respective scales were Clans (.72), Market (.68), and Hierarchy (.67).

There is some controversy about the measurement of organizational culture (Zammuto and Krakower, 1991). For example, some researchers have suggested that quantitative measures of culture are not appropriate to the measure of group social behavior because it is “too much a product of the social scientist’s rather than participant’s point of view” (Ouchi and Wilkins, 1985: 470). This is countered by oth-ers who believe that “the whole point of the contemporary study of organizational culture is to go beyond the method of the anthropologist by applying multivariate statistical analyses” (Ouchi and Wilkins: 478). This study utilized Likert measurement scales as suggested by researchers when correlation-based analyses (e.g., factor analysis, regression analysis) are performed (Quinn and Spreitzer, 1991). The Likert items assessed institutional character, leadership style, institutional cohesion, and institutional emphases. The items describe the core values and orientation of the institution.

Strategic orientation was assessed by Likert items that measured institutional behaviors and characteristics. These items were derived from research by Miles and Snow (1978). The two strategic types and their respective scales were ac-tive (.79) and passive (.70). An active strategy can be char-acterized as innovative and risk taking. A passive strategy can be described as insular, reactive, and risk averse.

The seven dimensions of organizational effectiveness were based on the Cameron (1978) effectiveness construct that examines perceived effectiveness of various domains of colleges and universities. The domains used in this study are student educational satisfaction (SES) (.71), ability to acquire resources (RES) (.72), system openness and com-munity interaction (OPE) (.70), student career development (SCD) (.61), faculty employment satisfaction (FES) (.62), professional development and quality of faculty (PDF) (.81), and organizational health (OH) (.76). The variables, their definitions and coefficient alphas are listed in Table 1.

Results. A 3x2-multivariate analysis of variance (MANOVA) design was used to analyze the data. The two independent variables were culture type (clan, market, hier-archy) and management strategy (active or passive). The dependent variables were the seven dimensions of organi-zational effectiveness.


Multivariate tests were significant for both main effects (culture: Wilk’s lambda =. 84, F = 2.83, p < .001, and man-agement strategy: Wilk’s lambda = .85, F = 5.14, p < .001). However, there were no significant interactive effects be-tween the two independent variables (see Table 2). Signifi-cant differences were established between each of the two independent variables and the dependent variables. Clan cultures were found to be more effective on three of the ef-fectiveness dimensions than market or hierarchy cultures

(System Openness and Community Interaction, Student Edu-cational Satisfaction, Faculty Employment Satisfaction, see Table 3). Active management strategies were found to be more effective than passive management strategies on six of the seven dimensions (Organizational Health, Ability to Acquire Resources, System Openness and Community In-teraction, Student Educational Satisfaction, Student Career Development, Faculty Employment Satisfaction, see Table 4), regardless of type.

Independent Variables Definition Coefficient Alpha

Organizational CultureClan

Emphasizes shared values and goals,participitativeness, sense of family.

0.72

Market Emphasizes competitiveness, goal accomplishment andproduction, consumer oriented.

0.68

Hierarchy Emphasizes rules and regulations, clear lines ofauthority, efficient and directive.

0.67

Managerial Strategy Active Innovative, continually developing, willing to break

precedents.0.79

Passive Insular, unable to strategize for the future, infrequentchange, no focused method to evaluate change.

0.70

Dependent Variables Definition Coefficient Alpha

Student EducationalSatisfaction (SES)

The extent of student satisfaction with the educationalresponse.

0.71(3 items)

Ability to AcquireResources (RES)

The extent to which the institution acquires resourcesfrom the external environment including finances, highquality students and faculty, research support, andpolitical legitimacy.

0.72(5 Items)

Systems Openness andCommunity Interaction(OPE)

The emphasis placed on the interaction with, adaptationto, and service in the institution’ external environment.

0.70(4 items)

Student CareerDevelopment (SCD)

The extent of occupation or vocational development ofstudents and the opportunities for occupationaldevelopment provided by the institution.

0.61(3 items)

Faculty EmploymentSatisfaction (FAS)

The satisfaction of faculty members with theiremployment and jobs at the institution.

0.62(3 items)

Professional Developmentand Quality of Faculty(PDF)

The extent of professional attainment and developmentof the faculty, and the a of stimulation towardprofessional development provided by the institution.

0.81(4 items)

Organizational Health(OH)

The extent of smooth functioning of the institution interms of its processes and operations, including thebenevolence and vitality of the institution.

0.76(4 items)

Table 1 Definition of Variables and Corresponding Alpha Coefficients


Conclusion

These findings provide evidence for differentiation of the three types of internal organizational culture as hypoth-esized by Ouchi (1980). Clan and market cultures comprise approximately half of the institutions surveyed with hierar-chical institutions comprising the remainder. Clan cultures were found to be more effective, as expected, on the dimen-sions that focus on the individual (Student Educational Sat-isfaction and Faculty Employment Satisfaction) and on the perceived “openness” of the school of business (System Openness and Community Interaction). Clan cultures can be described as having a sense of family, with a common goal. However, no statistically significant differences were found between the three culture types on the other four di-

mensions of effectiveness, suggesting that for business schools, the effectiveness dimensions that are more exter-nally focused are not affected by the culture type of the or-ganization (e.g., Professional Development and Quality of Faculty, Organizational Health, Ability to Acquire Re-sources, and Student Career Development).

We suggest several possible explanations for this lack of effect. First, the lack of effect of culture type on the ex-ternally focused effectiveness domains may be a result of the external environment. Today’s environment for higher education is particularly tumultuous. The public is demand-ing more accountability; economic stresses upon institutions are increasing, and changing student demographics are ef-fecting the environment in a number of unpredictable ways. These factors may be suppressing the impact of culture type on the externally focused effectiveness domains (Shein, 1996). Secondly, while the data suggested the primary cul-ture type, it did not indicate the “strength” of the culture. Previous research has suggested that an alignment between the beliefs and values central to the organization and actual strategies is the central distinguishing feature of strong cul-tures (Dennison, 1990). There may be subordinate cultures operating at the department level that flourish within the larger institutional culture, but do not necessarily align with the primary culture’s beliefs and values (Wilkins and Ouchi, 1983). Subordinate cultures may negate the effect of the primary culture. Finally, the lack of effect of culture type on the externally focused domains may be due to the fact that organizational members of business schools have unique values and orientation due to the nature of the disciplines (Cheit, 1985; Hugstad, 1983). Members of professional schools are aligned to their academic department and disci-pline and often to their professional practice (Clark, 1989). Professional schools expect their faculty to keep current with the profession through active practice. There may be a ten-sion between the “academic world” and the “applied world” that weakens the overall culture of the organization. Ap-plied environments are focused on short-term results, rapid decisions, teamwork, and practical application of theory and knowledge. Academic environments on the other hand, are typically slow and deliberate in their decision making, fo-cus on the long-term, and stress the abstract. While profes-sional school faculties intentionally attempt to bridge the gap between academics and practice, that tension may di-minish the overall effectiveness of the organization.

The study essentially corroborates previous research suggesting that managerial strategy is a strong indicator of organizational effectiveness (Child, 1972; Cameron, 1983). The active management strategy was found to be more ef-fective on six of the seven dimensions of effectiveness re-gardless of culture type. The results suggest that management strategy works independently of culture type, and that re-gardless of culture type, leaders should strive to be active in their strategies rather than passive. Additional research should address the differences between graduate and under-graduate programs. No effort was made in this study to sepa-

Table 2 Culture by Management Manova of Seven Dependent Variables Source of Variation df Wilks Lambda F p

Culture 14 .84 2.83 .001 Management Strategy 7 .85 5.14 .001 Culture x Management Strategy 14 .92 1.23 .246 Error 14 .93 1.03 .422

Table 3 Effectiveness Scale Score by Culture Type

Clan Market Hierarchy

Professional Development and Quality of Faculty 19.55 20.45 18.81 Organizational Health 20.48 19.47 20.24 Ability to Acquire Resources 17.48 18.85 16.94 Systems Openness and Community Interaction 16.95* 13.18 14.63 Student Educational Satisfaction 13.57* 11.81 12.52 Student Career Development 16.34 16.49 15.74 Faculty and Administrator Employment Satisfaction 15.21* 12.56 14.13 *p<0.05

Table 4 Effectiveness Scale Score by Managerial Strategy

Active Passive

Professional Development and Quality of Faculty 20.32 18.89 Organizational Health 21.84* 18.29 Ability to Acquire Resources 19.12* 16.40 System Openness and Community Interaction 16.35 13.49 Student Educational Satisfaction 13.22* 12.05 Student Career Development 16.76* 15.61 Faculty and Administrator Employment Satisfaction 15.36* 12.56 *p<0.05


rate graduate business programs from undergraduate busi-ness programs but rather to address the nature of business schools in total. Are there differences in perceptions as re-lated to the existence of graduate programs alongside un-dergraduate programs? The great bulk of discussion on business schools in the popular and business press centers around Masters in Business Administration (MBA) pro-grams. It is unclear if the presence of graduate programs differentially effect any of the institutional effectiveness outcomes examined in this study.

The study is limited by its reliance on one respondent per school. A richer picture of the institution would have been developed had there been several responses from each school. A larger number of responses that included faculty members would have also allowed us to gauge the strength of the primary culture type. However, the study begins to build an understanding of the uniqueness of professional schools and the dilemmas they face that may be different from higher education in general. As more and more stu-dents enter higher education specifically to gain a profes-sional expertise and gainful employment, understanding the special characteristics of professional schools is essential in measuring organizational effectiveness.

References

Ansoff, H. I. and Hayes, R. L. (1976). Introduction. In H. I.Ansoff, R. P. Declerck, and R. L. Hayes (Eds.), From Strategic Planning to Strategic Management, (pp.3-6). New York: Wiley.

Ansoff, Arnold D. R. and Capella, L. M. (1985). Corporate culture and the marketing concept: a diagnostic instru-ment for utilities. Public Utilities Fortnightly. 116, 32- 38.

Axinn, C. N. and Thach, S. V. (1996). Reflections on gradu-ate education in the 21st century. Mid-American Journal of Business, 11, 5-9.

Cameron, K. S. (1978). Measuring organizational effective-ness in institutions of higher education. Administrative Science Quarterly, 23, 604-632.

Cameron, K. S. (1986). A study of organizational effective-ness and its predictors. Management Science, 32, 87- 111.

Cameron, K. S. and Ettington, D. R. (1988). Conceptual foundations of organizational culture. In J. C. Smart (Ed.). Higher Education: Handbook of Theory and Research, Vol. 4, (pp.356-396). New York: Agathon Press.

Cameron, K. S. (1987). Measuring organizational effective-ness in institutions of higher education. Administrative Science Quarterly, 23, 604-632.

Cameron, K. S. (1983). Strategic responses to conditions of decline: Higher education and the private sector. Jour-nal of Higher Education, 54, 359-380.

Cameron, K. S. and Tischirhart, M. (1992). Postindustrial environments and organizational effectiveness in colleges and universities. Journal of Higher Education, 63, 87- 108.

Cameron, K. S. and Zammuto, R. (1983). Matching mana-gerial strategies to conditions of decline. Human Re-source Management, 22, 359-375.

Chaffee, E. E. (1986). The concept of strategy: From busi-ness to higher education. In J. C. Smart (Ed.). Higher Education: Handbook of Theory and Research, Vol. 1, (pp.133-172). New York: Agathon Press.

Chandler, A. D.(1962). Strategy and Structure. Cambridge, MA: MIT Press.

Cheit, E. F.(1985). Business schools and their critics. Cali-fornia Management Review, 27, 43-62.

Child, J. (1972). Organization structure, environment and performance: The role of strategic choice. Sociology, 6, 1-22.

Clark, B. R.(1989). The academic life-small worlds, differ-ent worlds. Educational Researcher, 18, 4-8.

Clott, C. B. (1994). Predictors of organizational effective-ness in schools of business. Unpublished doctoral dis-sertation. University of Illinois at Chicago, Chicago, IL.

Davis, S. and Schwartz, H. (1981). Matching Corporate Culture and Business Strategy. Concept Paper Series, Cambridge, MA: Management Analysis Center.

Deal, T. E. and Kennedy, A. A. (1982). Corporate Cultures. Reading, MA: Addison-Wesley.

Denison, D. R. and Spreitzer, G. M. (1991). Organizational culture and organizational development. In L. L. Cummings and B. M. Staw (Eds.), Research in Organi-zational Change and Development, Vol. 5 (pp.1-21). Greenwich, CT: JAI Press Inc.

Doty, D. H., Glick, W. H., and Huber, G. P. (1993). Fit, equifinality, and organizational effectiveness: A test of two configurational theories. Academy of Management Journal, 36,1196-1250.

Hambrick, D. C. (1983). Some tests of the effectiveness and functional attributes of Miles and Snow’s strategic types. Academy of Management Journal, 26, 2-26.

Hatten, K. J. (1979). Quantitative research methods in stra-tegic management. In D. E. Schendel and C. W. Hofer (Eds.), Strategic Management, Boston: Little Brown.

Hugsatd, P. S. (1983). The Business School in the 1980s. New York: Praeger Publishers.

Katz, D. and Kahn, R. L. (1978). The Social Psychology of Organizations, 2nd ed. New York: Wiley.

Miles, R. H. and Cameron, K. S. (1982). Coffin Nails and Corporate Strategies. Englewood Cliffs, NJ: Prentice- Hall.


Miles, R. E. and Snow, C. E. (1978). Organizational Strat-egy, Structure, and Process. New York: McGraw-Hill Book Co.

Ouchi, W. G. (1980). Markets, bureaucracies, and clans. Administrative Science Quarterly, 25, 129-141.

Ouchi, W.G. (1981). Theory Z: How American Business Can Meet the Japanese Challenge. Reading MA: Addison- Wesley.

Parsons, T. (1951). The Social System. New York: Free Press. Pennings, J. M. and Goodman, P. S. (1977). New Perspec-

tives on Organizational Effectiveness. San Francisco: Jossey-Bass.

Peters, T. and Waterman, R. (1982). In Search of Excellence. New York: Harper and Row.

Pettigrew, A. M. (1979). On studying organizational cul-tures. Administrative Science Quarterly, 24, 570-581.

Quinn, R.E. (1990). Beyond Rational Management. San Francisco: Jossey-Bass.

Quinn, R. E. and Rohrbaugh, J. (1983). A spatial model of effectiveness criteria: Towards a competing values ap-proach to organizational analysis. Management Science, 29, 363-377.

Schein, E. H. (1985). Organizational Culture and Leader-ship. San Francisco: Jossey-Bass.

Schein, E. H. (1992). Organizational Culture and Leader-ship (2nd ed.). San Francisco: Jossey-Bass.

Shein, E. H. (1996). Culture: The missing concept in orga-nization studies. Administrative Science Quarterly, 41, 229-240.

Snow, C. C. and Hrebiniak, L. G. (1980). Strategy, distinc-tive competence, and organizational performance. Admin-istrative Science Quarterly, 25, 317-336.

Swidler, A. (1986). Culture in action. American Sociologi-cal Review, 51, 273-287.

Tichy, N. M. (1983). Managing Strategic Change. New York: John Wiley and Sons.

Wilkins, A. L. and Ouchi, W. G. (1983). Efficient cultures: Exploring the relationship between culture and organi-zational performance. Administrative Science Quarterly, 28, 468-481.

Yuchtman, E. and Seashore, S. (1967). A system resource approach to organizational effectiveness. American So-ciological Review, 32, 891-903.

Zammuto, R. F. and Krakower, J. Y. (1991). Quantitative and qualitative studies of organizational culture. In L. L. Cummings and B. M. Staw (Eds.) Research in Organiza-tional Change and Development. Vol. 5, (pp. 83-114). Greenwich, CT:JAI Press Inc.

Mid-Western Educational Researcher Call for Manuscripts

The Mid-Western Educational Researcher is a scholarly journal that publishes research-based articles addressing a full range of educational issues. The journal also publishes literature reviews, theoretical and methodological discussions that make an original contribution to the research literature, and feature columns. There are four issues of the journal published annually.

The journal is now accepting manuscripts for review and possible publication in 2000 and beyond. Manuscripts are submitted to blind reviews by at least two researchers with knowledge of the literature in the appropriate area. Furthermore, the editors will review the manuscript and make the final decision. The review process requires approximately three months.

Manuscripts are accepted from faculty, students, and professionals working in non-educational settings. Membership in the MWERA is not required in order to submit a manuscript for review. The editors encourage the submission of revised papers that have been presented at the annual meetings of the MWERA, AERA, and other professional organizations.

Submit four (4) copies of the manuscript with a cover letter to Mary K. Bendixen-Noe (see address below). Manu-scripts should conform to the style and format described in the Publication Manual of the American Psychological Asso-ciation, 4th edition. All manuscripts should be typed, double spaced, on 8 1/2 x 11 paper with 1 1/2 inch margins on all sides. An abstract of less than 100 words should accompany the manuscript. The author’s name and affiliation should appear on the title page only. Submissions typically are less than 20 pages in length. If the manuscript is accepted for publication, the author(s) will be asked to provide a disk file (WP 5.1 or higher on a 3 1/2 inch high density IBM formatted disk) as well as a printed copy of the final version. Please note that the editors reserve the right to make minor modifica-tions in order to produce a concise and clear article.

Questions regarding the journal or the submission of feature columns should be directed to the editor.

Mary K. Bendixen-Noe The Ohio State University, Newark 1179 University Dr. Newark, OH 43055 Phone: (740) 366-9469 Fax: (740) 366-5047 email: [email protected]


Introduction

Educational researchers are often interested in making a statistical inference regarding a population correlation coefficient (Pearson product-moment) when they have data for a sample from that population. The most common ap-proach is to test the null hypothesis that in the population there is no linear relationship between two variables X and Y, so they carry out a t test of the significance of the sample correlation coefficient, using the formula t = r/[(1-r2)/(n-2)].5, for n-2 degrees of freedom, where r is the sample correla-tion and n is the sample size (see, for example, Agresti and Finlay, 1986, pp. 272-274). Testing some hypothesis other than 0 or constructing a confidence interval around the sample r is a bit more complicated, because both require Fisher’s r-to-z transformation.

But what about the underlying assumptions? Is vari-able X “fixed” or “random”? If it is fixed, is Y normal and homoscedastic? If random, are X and Y bivariate normal? Those who are concerned about such assumptions have re-cently been attracted to the nonparametric “resampling” approach to statistical inference, where such assumptions are relaxed or not necessary.

This article compares and contrasts the traditional para-metric approach with the two most popular resampling ap-proaches, “the jackknife” and “the bootstrap”, using an example involving the relationship between LSAT score and GPA for 15 law schools. This example was first introduced by Efron (1979) and has been used by several authors (see, for example, Efron, 1982; Diaconis and Efron, 1983; Efron and Gong, 1983, Efron and Tibshirani, 1993, and Lee and Rodgers, 1998).

Methods

The example

The data are as follows (note that the unit of analysis is the law school, not the individual student):

School Average Average Undergraduate LSAT (X) GPA (Y)

1 576 3.39 2 635 3.30 3 558 2.81 4 578 3.03 5 666 3.44 6 580 3.07 7 555 3.00 8 661 3.43 9 651 3.36 10 605 3.13 11 653 3.12 12 575 2.74 13 545 2.76 14 572 2.88 15 594 2.96

The Traditional Approach

In the traditional approach the sample correlation coef-ficient would be calculated and tested for statistical signifi-cance and/or a confidence interval for the population correlation coefficient would be constructed.

“Jackknifing”1

About 50 years ago Quenouille (1949) developed a gen-eral nonparametric procedure for estimating sampling error when one is unable or unwilling to make the assumptions that are necessary for the traditional parametric approach to statistical inference. This method was later popularized by Tukey (see Tukey, 1958; Mosteller and Tukey, 1977; and elsewhere). It proceeds as follows: 1. Calculate the sample statistic for the entire sample of

size n; call that yall (where Y is the dependent variable). 2. Calculate n jackknifed values y(j) for that statistic by

deleting from the sample data a different observation each time. (A variation is sometimes employed whereby more than one observation is deleted each time.) This “sampling of the sample” treats the sample observa-tions as miniature representations of the unknown popu-lation observations.

Traditional vs. “Resampling” Approaches to Statistical Inferences Regarding Correlation Coefficients

Thomas R. Knapp, Gerald L. Noblitt, and Sunanta Viragoontavan The Ohio State University

Abstract With the ever-increasing use of computers in data analysis, there has been a movement to abandon tradi-tional parametric approaches, with all of their restrictive assumptions, in favor of computer-intensive nonparametric inferential statistical procedures such as the jackknife and the bootstrap that are based upon resampling of the sample data. In this article these techniques are compared and contrasted with the parametric approach, with special emphasis on inferences regarding correlation coefficients.


Table 2 Sampling distribution of 1000 bootstrapped correlations.

r Frequency % Cumulative %

below .260 2 .20 .20 .260 to .299 1 .10 .30 .300 to .339 2 .20 .50 .340 to .379 2 .20 .70 .380 to .419 7 .70 1.40 .420 to .459 8 .80 2.20 .460 to .499 14 1.40 3.60 .500 to .539 13 1.30 4.90 .540 to .579 28 2.80 7.70 .580 to .619 42 4.20 11.90 .620 to .659 58 5.80 17.70 .660 to .699 75 7.50 25.20 .700 to .739 95 9.50 34.70 .740 to .779 131 13.10 47.80 .780 to .819 117 11.70 59.50 .820 to .859 104 10.40 69.90 .860 to .899 125 12.50 82.40 .900 to .939 121 12.10 94.50 .940 to .979 53 5.30 99.80 above .979 2 .20 100.00 mean = .774 (a slight bias) estimated standard error = .128

3. Calculate n “pseudovalues” y*j for that same statistic. These pseudovalues are not necessarily on the same scale as the original statistic, but they are a set of num-bers whose sampling variability is on a par with what would be obtained by random sampling of the popula-tion. (The sampling variability of the y(j) is too small, since they only differ from the original statistic by us-ing just one fewer observation). The formula for the pseudovalues is: y*j = n(yall) - (n-1)(y(j)).

4. Calculate the mean of the pseudovalues and the standard error of that mean, and use the standard error to carry out a significance test and/or to construct a confidence interval.

“Bootstrapping”2

Unlike the jackknife for which the resampling is without replacement (where each jackknifed statistic is based upon n- 1 of the n observations), the bootstrap approach tries to resolve the inference dilemma by “sampling the sample” with replace-ment (where each bootstrapped statistic is based upon n obser-vations). It usually accomplishes this by a simulation process (you can’t properly bootstrap without a computer), as follows: 1. Copy the sample data a large number of times (Diaconis

and Efron, 1983, give the figure of a billion times!) 2. Sample n observations from this “population” k times

(where k is also very large). 3. Calculate the statistic of interest for each sample and gen-

erate the empirical sampling distribution for that statistic. 4. Calculate the standard deviation (standard error) of that

sampling distribution and make the appropriate inference.

Results

The Traditional Approach

The sample r is .776 and Fisher’s z = 1.035. For n = 15 that correlation is statistically significant at the .05 level and the 95% confidence interval for the population correlation extends from .469 to 1.601 on the Fisher’s z scale and from .437 to .922 on the r scale (again see Agresti and Finlay, 1986, pp. 272-274 for details).

Jackknifing

The jackknifed statistics, pseudovalues, mean pseudovalue, and standard error are displayed in Table 1. Since the estimated standard error is .143, an appeal to ro-bustness or to Chebyshev’s inequality would suggest that the r of .776 is significant at least at .05, and the 95% confi-dence interval would extend from about .496 to about .999. (The upper limit is actually greater than 1, but it is conven-tional to cut it off just below 1.)

Bootstrapping

For bootstrapping of these data, 1000 samples of size 15 were drawn from that “population” of bivariate observa-

tions. The empirical sampling distribution of the 1000 bootstrapped correlations is displayed in Table 2. Those correlations ranged from .251 to .990, with a mean of .774 and a standard deviation (standard error) of .128. The dis-tribution was highly skewed. The middle 95% of the distri-bution extended from about .460 to about .960, so the null hypothesis that the population correlation is zero would be rejected, and one could be “95% confident” that the interval from .460 to .960 would “cover” the true value of that popu-lation correlation.

Table 1 Jackknife results for the law schools data. (r = yall = .776)

omitted observation j y(j) pseudovalue y*j

1 .893 -.862 2 .764 .944 3 .755 1.070 4 .776 .776 5 .731 1.406 6 .780 .720 7 .785 .650 8 .736 1.336 9 .752 1.112 10 .776 .776 11 .818 .118 12 .786 .636 13 .740 1.280 14 .767 .902 15 .780 .720

mean pseudovalue = .772 (a slight bias) estimated standard error = .144


Discussion

For this example all three approaches yielded similar results, with slightly different standard errors. For other data sets the results might not agree nearly as well.

The most obvious advantage over the traditional ap-proach for both jackknife and bootstrap is that there is no need to assume population normality or homoscedasticity. The most obvious disadvantage is the greater amount of computational time necessary for either of those resampling approaches.

Perhaps a more interesting comparison is between jack-knife and bootstrap. The bootstrap technique is somewhat easier to understand, primarily because the jackknife “pseudovalues” are unfamiliar. (Since they do not have to be on the same scale as the statistic of interest, the pseudovalues for correlation coefficients are often less than -1 and greater than +1.) And the mathematical statisticians claim that the bootstrap has some important technical ad-vantages (for example, it appears to work better for non- linear statistics such as the sample median--see Mooney and Duval, 1993). One problem, however, is that in boot-strap sampling you may draw the same observation every time. It was at least theoretically possible, although highly improbable, that Law School #1, for example (which is ac-tually an outlier), could have gotten sampled 15 times in one of the samples of 15 observations, rendering that sample atypical of the original data and making it impossible to even calculate its Pearson r, since there would be no variability in either X or Y. (See Wainer and Thissen, 1975 for an ex-ample of the effect of an outlier on the jackknife approach.) In this application the “worst” samples in that respect were one sample in which School #14 was drawn six times and another sample in which School #15 was drawn six times.

The jackknife may be easier for most people to imple-ment, and it avoids the problem of non-independence of the observations when the same unit is sampled more than once. It has the additional advantage that you don’t have to trust random (actually pseudorandom) number generators. (See Whitney, 1984 for a discussion of problems with pseudo-random numbers.)

Some Final Comments

There is nothing special about correlation coefficients. Jackknifing and bootstrapping have been applied to means, standard deviations, percentiles (e.g., medians), R-squares, and many other statistics, with very encouraging results. (See Fan and Wang, 1996 for an interesting comparison between bootstrap and jackknife alternatives to canonical correlation analysis.) However, resampling approaches must not be oversold the way that non-parametric statistics in general were oversold in the 1950s and 1960s. If for a given situa-tion parametric assumptions such as normality and

homoscedasticity are met or “near-met” (by an appeal to robustness) it is a waste of time and effort, to say nothing about a potential lack of power, to jackknife or bootstrap.

Footnotes

1 The term “jackknife” is due to Tukey. He used the analogy to an ordinary jackknife, which is a very useful tool for many purposes.

2 The term “bootstrap” is due to Efron. When people get into trouble, e.g., in statistical inference, they shouldn’t give up; they need to pull themselves up by their bootstraps.

References

Agresti, A. and Finlay, B. (1986). Statistical methods for the social sciences (2nd. ed.). San Francisco: Dellen.

Diaconis, P. and Efron, B. (1983). Computer-intensive meth-ods in statistics. Scientific American, 116-130.

Efron, B. (1979). Computers and the theory of statistics: Thinking the unthinkable. SIAM Review, 21, 460-480.

Efron, B. (1982). The jackknife, the bootstrap, and other resampling plans. Philadelphia: Society for Industrial and Applied Mathematics.

Efron, B. and Gong, G. (1983). A leisurely look at the boot-strap, the jackknife, and cross-validation. The American Statistician, 37, 36-48.

Efron, B. and Tibshirani, R. J. (1993). An introduction to the bootstrap. New York: Chapman and Hall.

Fan, X. and Wang, L. (1996). Comparability of jackknife and bootstrap results: An investigation for a case of ca-nonical correlation analysis. The Journal of Experimen-tal Education, 64(2), 173-189.

Lee, W-C and Rodgers, J. L. (1998). Bootstrapping corre-lation coefficients using univariate and bivariate sampling. Psychological Methods, 3, 91-103.

Mooney, C. Z. and Duval, R. D. (1993). Bootstrapping: A nonparametric approach to statistical inference. Newbury Park, CA: Sage.

Mosteller, F. and Tukey, J. W. (1977). Data analysis and regression. Reading, MA: Addison-Wesley.

Quenouille, M. (1949). Approximate tests of correlation in time series. Journal of the Royal Statistical Society, Se-ries B, 11, 18-84.

Tukey, J. W. (1958). Bias and confidence in not quite large samples. Annals of Mathematical Statistics, 29, 614.

Wainer, H. and Thissen, D. (1975). When jackknifing fails (or does it?). Psychometrika, 40, 113-114.

Whitney, C. A. (1984). Generating and testing pseudoran-dom numbers. Byte, 9, 128-129, 442, 444, 446, 449- 450, 452, 456-459, 462-464.

The Mid-Western Educational Research Association Gift Membership

A gift membership has been given to you,

by

Your name is now included as a member in one of the most recognized, well re-spected, educational research groups in the United States and Canada. Your one year membership includes a subscription to the Mid-Western Educational Re-searcher, the Association’s journal that highlights research articles, features, in-terviews, and Association news. Members pay reduced registration fees for the annual meeting held in Chicago in October. This conference attracts many nation-ally recognized leaders in educational research. Enjoy your membership.

Thank you for providing your colleague, student, or friend with a special one year gift membership to the Mid-Western Educational Research Association. It is a gift of professional involvement that is sure to be appreciated throughout the year. To give your gift membership fill out the top portion of this card and use it to inform the recipient of the gift membership; then fill out the bottom portion of this card and mail it with your check to: Jean W. Pierce - Dept EDCSE - Northern Illinois Univ. - DeKalb, IL 60115

Check one below and make check payable to Mid-Western Educational Research Association.

Professional Membership - $25 Student Membership - $15 Student must be currently enrolled.

Name _____________________________ _____________________________

Address _____________________________ _____________________________

_____________________________ _____________________________

_____________________________ _____________________________

Affiliation _____________________________ _____________________________

Work Phone _____________________________ _____________________________

Home Phone _____________________________ _____________________________

E-mail _____________________________ _____________________________

Fax _____________________________ _____________________________

Division _____________________________ _____________________________

Person Receiving Gift Membership Person Giving Gift Membership

Preference (optional)

The Ohio State University 2500 Kenny Road Columbus, Ohio 43210

Volume 13, No. 2 Spring 2000 • Official Publication of the ...ies evolving from theoretical models...

Documents

Transcript of Volume 13, No. 2 Spring 2000 • Official Publication of the ...ies evolving from theoretical models...